This document proposes a method for automatically detecting compound structures from multiple hierarchical image segmentations. Compound structures consist of spatial arrangements of primitive objects. The method models compound structures as probabilistic region processes and learns their characteristics. Candidate regions are extracted from hierarchical segmentations and the most coherent subsets are selected that constitute compound structure instances. The selection is formulated as optimizing a constrained region selection framework.
PROPOSED ALGORITHM OF ELIMINATING REDUNDANT DATA OF LASER SCANNING FOR CREATI...IAEME Publication
Aerial laser scanning for creating digital relief models has been widely applied in
Russia for over 20 years. For processing results of aerial laser scanning, it is
necessary to classify cloud of laser points. Classification of cloud of laser points is
performed via commercial software products (produced abroad), which frequently use
refinement algorithms of “Earth” class data, which consider the relief particularities.
The paper puts forward a proprietary algorithm of laser scanning data interpolation,
enabling to remove redundant cloud of laser points of “Earth” class when data
granularity reduces, in the first place, for flat terrain. The paper provides a detailed
stepwise description of algorithm operation, and the results of constructing a digital
relief model on the basis of the cloud of laser points of “Earth” class, processed via
the proposed algorithm.
The retrieval algorithms in remote sensing generally involve complex physical forward models that are nonlinear and computationally expensive to evaluate. Statistical emulation provides an alternative with cheap computation and can be used to calibrate model parameters and to improve computational efficiency of the retrieval algorithms. We introduce a framework of combining dimension reduction of input and output spaces and Gaussian process emulation
technique. The functional principal component analysis (FPCA) is chosen to reduce to the output space of thousands of dimensions by orders of magnitude. In addition, instead of making restrictive assumptions regarding the correlation structure of the high-dimensional input space,
we identity and exploit the most important directions of this space and thus construct a Gaussian process emulator with feasible computation. We will present preliminary results obtained from applying our method to OCO-2 data, and discuss how our framework can be generalized in
distributed systems. This is joint work with Jon Hobbs, Alex Konomi, Pulong Ma, and Anirban Mondal, and Joon Jin Song.
PROPOSED ALGORITHM OF ELIMINATING REDUNDANT DATA OF LASER SCANNING FOR CREATI...IAEME Publication
Aerial laser scanning for creating digital relief models has been widely applied in
Russia for over 20 years. For processing results of aerial laser scanning, it is
necessary to classify cloud of laser points. Classification of cloud of laser points is
performed via commercial software products (produced abroad), which frequently use
refinement algorithms of “Earth” class data, which consider the relief particularities.
The paper puts forward a proprietary algorithm of laser scanning data interpolation,
enabling to remove redundant cloud of laser points of “Earth” class when data
granularity reduces, in the first place, for flat terrain. The paper provides a detailed
stepwise description of algorithm operation, and the results of constructing a digital
relief model on the basis of the cloud of laser points of “Earth” class, processed via
the proposed algorithm.
The retrieval algorithms in remote sensing generally involve complex physical forward models that are nonlinear and computationally expensive to evaluate. Statistical emulation provides an alternative with cheap computation and can be used to calibrate model parameters and to improve computational efficiency of the retrieval algorithms. We introduce a framework of combining dimension reduction of input and output spaces and Gaussian process emulation
technique. The functional principal component analysis (FPCA) is chosen to reduce to the output space of thousands of dimensions by orders of magnitude. In addition, instead of making restrictive assumptions regarding the correlation structure of the high-dimensional input space,
we identity and exploit the most important directions of this space and thus construct a Gaussian process emulator with feasible computation. We will present preliminary results obtained from applying our method to OCO-2 data, and discuss how our framework can be generalized in
distributed systems. This is joint work with Jon Hobbs, Alex Konomi, Pulong Ma, and Anirban Mondal, and Joon Jin Song.
The retrieval algorithms in remote sensing generally involve complex physical forward models that are nonlinear and computationally expensive to evaluate. Statistical emulation provides an alternative with cheap computation and can be used to calibrate model parameters and to improve computational efficiency of the retrieval algorithms. We introduce a framework of combining dimension reduction of input and output spaces and Gaussian process emulation technique. The functional principal component analysis (FPCA) is chosen to reduce to the output space of thousands of dimensions by orders of magnitude. In addition, instead of making restrictive assumptions regarding the correlation structure of the high-dimensional input space, we identity and exploit the most important directions of this space and thus construct a Gaussian process emulator with feasible computation. We will present preliminary results obtained from applying our method to OCO-2 data, and discuss how our framework can be generalized in distributed systems.
Introduction Petrel Course (UAB-2014)
This course has been prepared as an introduction of Petrel software (Schlumberger, www.software.slb.com/products/platform/Pages/petrel.aspx), an application which allows the modeling and visualization of reservoirs, since the exploration stage until production, integrating geological and geophysical data, geological modeling (structural and stratigraphic frameworks), well planning, or property modeling ( petrophysical or petrological) among other possibilities.
The course will be focused mainly in the understanding and utilization of workflows aimed to build geological models based on superficial data (at the outcrop scale) but also with seismic data. The course contents have been subdivided in 5 modules each one developed through the combination of short explanations and practical exercises.
The duration of the course covers more or less 10h divided in three sessions. The starting data will be in the first week of December.
This course will be oriented mainly for the PhD and master students ascribed at the Geologic department of the UAB. For logistic reasons the maximum number of places for each torn are 9. The course is free from the Department members but the external interested will have to make a symbolic payment.
Those interested send an e-mail to the Doctor Griera (albert.griera@uab.cat).
The course will be imparted by Marc Diviu (Msc. Geology and Geophysics of reservoirs).
4-CONNECTED AND 8-CONNECTED NEIGHBOR SELECTION By Sintiak HaqueSintiak haque
Boundary fill algorithm is used frequently in computer graphics to fill a desired color inside a closed polygon having the same boundary color for all of its sides.Boundary Fill Algorithm starts at a pixel inside the polygon to be filled and paints the interior proceeding outwards towards the boundary. This algorithm works only if the color with which the region has to be filled and the color of the boundary of the region are different.
User guide of reservoir geological modeling v2.2.0Bo Sun
This is the user guide of DepthInsight™ reservoir geological modeling module. For corresponding video tutorials , please visit and subscribe our Youtube channel: https://www.youtube.com/channel/UCjHyG-mG7NQofUWTZgpBT2w
DepthInsight™ software products include modules as follows:
Structure Interpretation
Well and Data Management
Plan Module
Profile Module
Attribute Modeling
Velocity Modeling
Structural Modeling
Reservoir Geological Modeling
Numerical Simulation Gridding
Rock Modeling
Geo-mechanical Modeling
Paleo-Structural Modeling
Enormous Modeling Platform
For more information about our company, Beijing GridWorld Software Technology Co., Ltd., please visit our website: http://gridworld.com.cn/en/
This presentation includes:
Velocity modeling the principles and pitfalls
Well and seismic velocity data
Incorporating velocity data to build a reliable model in Petrel software
Time to Depth conversion (Map and reservoir property)
Residual error correction and well marker adjustment
Structural uncertainty
Monitoring traffic in urban areas is an important task for intelligent transport applications to alleviate the traffic problems like traffic jams and long trip times. The traffic flow in urban areas is more complicated than the traffic flow in highway, due to the slow movement of vehicles and crowded traffic flows in urban areas. In this paper, a vehicle detection and classification system at intersections is proposed. The system consists of three main phases: vehicle detection, vehicle tracking and vehicle classification. In the vehicle detection, the background subtraction is utilized to detect the moving vehicles by employing mixture of Gaussians (MoGs) algorithm, and then the removal shadow algorithm is developed to improve the detection phase and eliminate the undesired detected region (shadows). After the vehicle detection phase, the vehicles are tracked until they reach the classification line. Then the vehicle dimensions are utilized to classify the vehicles into three classes (cars, bikes, and trucks). In this system, there are three counters; one counter for each class. When the vehicle is classified to a specific class, the class counter is incremented by one. The counting results can be used to estimate the traffic density at intersections, and adjust the timing of traffic light for the next light cycle. The system is applied to videos obtained by stationary cameras. The results obtained demonstrate the robustness and accuracy of the proposed system.
International Journal of Engineering Research and DevelopmentIJERD Editor
Electrical, Electronics and Computer Engineering,
Information Engineering and Technology,
Mechanical, Industrial and Manufacturing Engineering,
Automation and Mechatronics Engineering,
Material and Chemical Engineering,
Civil and Architecture Engineering,
Biotechnology and Bio Engineering,
Environmental Engineering,
Petroleum and Mining Engineering,
Marine and Agriculture engineering,
Aerospace Engineering.
Surveillance refers to the task of observing a scene, often for lengthy periods in search of particular objects or particular behaviour. This task has many applications, foremost among them is security (monitoring for undesirable behaviour such as theft or vandalism), but increasing numbers of others in areas such as agriculture also exist. Historically, closed circuit TV (CCTV) surveillance has been mundane and labour Intensive, involving personnel scanning multiple screens, but the advent of reasonably priced fast hardware means that automatic surveillance is becoming a realistic task to attempt in real time. Several attempts at this are underway.
Special Plenary Lecture at the International Conference on VIBRATION ENGINEERING AND TECHNOLOGY OF MACHINERY (VETOMAC), Lisbon, Portugal, September 10 - 13, 2018
http://www.conf.pt/index.php/v-speakers
Propagation of uncertainties in complex engineering dynamical systems is receiving increasing attention. When uncertainties are taken into account, the equations of motion of discretised dynamical systems can be expressed by coupled ordinary differential equations with stochastic coefficients. The computational cost for the solution of such a system mainly depends on the number of degrees of freedom and number of random variables. Among various numerical methods developed for such systems, the polynomial chaos based Galerkin projection approach shows significant promise because it is more accurate compared to the classical perturbation based methods and computationally more efficient compared to the Monte Carlo simulation based methods. However, the computational cost increases significantly with the number of random variables and the results tend to become less accurate for a longer length of time. In this talk novel approaches will be discussed to address these issues. Reduced-order Galerkin projection schemes in the frequency domain will be discussed to address the problem of a large number of random variables. Practical examples will be given to illustrate the application of the proposed Galerkin projection techniques.
Video surveillance is becoming more and more important forsocial security, law enforcement, social order,military, and other social problems. In order to manage parking information effectively, this vehicle
detection method is presented. In general, motion detection plays an important role in video surveillance
systems. In this paper, firstly this system uses ViBe method to extract the foreground object, then extracts
HOG features on the performance of the ROI of images. At last this paper presents Support vector machine for vehicle recognition. The results of this test show that, the recognition rate of vehicle’s model in this recognition system is up the industrial application standard.
Video surveillance is becoming more and more important forsocial security, law enforcement, social order,military, and other social problems. In order to manage parking information effectively, this vehicle detection method is presented. In general, motion detection plays an important role in video surveillance systems. In this paper, firstly this system uses ViBe method to extract the foreground object, then extracts HOG features on the performance of the ROI of images. At last this paper presents Support vector machine for vehicle recognition. The results of this test show that, the recognition rate of vehicle’s model in this recognition system is up the industrial application standard.
The retrieval algorithms in remote sensing generally involve complex physical forward models that are nonlinear and computationally expensive to evaluate. Statistical emulation provides an alternative with cheap computation and can be used to calibrate model parameters and to improve computational efficiency of the retrieval algorithms. We introduce a framework of combining dimension reduction of input and output spaces and Gaussian process emulation technique. The functional principal component analysis (FPCA) is chosen to reduce to the output space of thousands of dimensions by orders of magnitude. In addition, instead of making restrictive assumptions regarding the correlation structure of the high-dimensional input space, we identity and exploit the most important directions of this space and thus construct a Gaussian process emulator with feasible computation. We will present preliminary results obtained from applying our method to OCO-2 data, and discuss how our framework can be generalized in distributed systems.
Introduction Petrel Course (UAB-2014)
This course has been prepared as an introduction of Petrel software (Schlumberger, www.software.slb.com/products/platform/Pages/petrel.aspx), an application which allows the modeling and visualization of reservoirs, since the exploration stage until production, integrating geological and geophysical data, geological modeling (structural and stratigraphic frameworks), well planning, or property modeling ( petrophysical or petrological) among other possibilities.
The course will be focused mainly in the understanding and utilization of workflows aimed to build geological models based on superficial data (at the outcrop scale) but also with seismic data. The course contents have been subdivided in 5 modules each one developed through the combination of short explanations and practical exercises.
The duration of the course covers more or less 10h divided in three sessions. The starting data will be in the first week of December.
This course will be oriented mainly for the PhD and master students ascribed at the Geologic department of the UAB. For logistic reasons the maximum number of places for each torn are 9. The course is free from the Department members but the external interested will have to make a symbolic payment.
Those interested send an e-mail to the Doctor Griera (albert.griera@uab.cat).
The course will be imparted by Marc Diviu (Msc. Geology and Geophysics of reservoirs).
4-CONNECTED AND 8-CONNECTED NEIGHBOR SELECTION By Sintiak HaqueSintiak haque
Boundary fill algorithm is used frequently in computer graphics to fill a desired color inside a closed polygon having the same boundary color for all of its sides.Boundary Fill Algorithm starts at a pixel inside the polygon to be filled and paints the interior proceeding outwards towards the boundary. This algorithm works only if the color with which the region has to be filled and the color of the boundary of the region are different.
User guide of reservoir geological modeling v2.2.0Bo Sun
This is the user guide of DepthInsight™ reservoir geological modeling module. For corresponding video tutorials , please visit and subscribe our Youtube channel: https://www.youtube.com/channel/UCjHyG-mG7NQofUWTZgpBT2w
DepthInsight™ software products include modules as follows:
Structure Interpretation
Well and Data Management
Plan Module
Profile Module
Attribute Modeling
Velocity Modeling
Structural Modeling
Reservoir Geological Modeling
Numerical Simulation Gridding
Rock Modeling
Geo-mechanical Modeling
Paleo-Structural Modeling
Enormous Modeling Platform
For more information about our company, Beijing GridWorld Software Technology Co., Ltd., please visit our website: http://gridworld.com.cn/en/
This presentation includes:
Velocity modeling the principles and pitfalls
Well and seismic velocity data
Incorporating velocity data to build a reliable model in Petrel software
Time to Depth conversion (Map and reservoir property)
Residual error correction and well marker adjustment
Structural uncertainty
Monitoring traffic in urban areas is an important task for intelligent transport applications to alleviate the traffic problems like traffic jams and long trip times. The traffic flow in urban areas is more complicated than the traffic flow in highway, due to the slow movement of vehicles and crowded traffic flows in urban areas. In this paper, a vehicle detection and classification system at intersections is proposed. The system consists of three main phases: vehicle detection, vehicle tracking and vehicle classification. In the vehicle detection, the background subtraction is utilized to detect the moving vehicles by employing mixture of Gaussians (MoGs) algorithm, and then the removal shadow algorithm is developed to improve the detection phase and eliminate the undesired detected region (shadows). After the vehicle detection phase, the vehicles are tracked until they reach the classification line. Then the vehicle dimensions are utilized to classify the vehicles into three classes (cars, bikes, and trucks). In this system, there are three counters; one counter for each class. When the vehicle is classified to a specific class, the class counter is incremented by one. The counting results can be used to estimate the traffic density at intersections, and adjust the timing of traffic light for the next light cycle. The system is applied to videos obtained by stationary cameras. The results obtained demonstrate the robustness and accuracy of the proposed system.
International Journal of Engineering Research and DevelopmentIJERD Editor
Electrical, Electronics and Computer Engineering,
Information Engineering and Technology,
Mechanical, Industrial and Manufacturing Engineering,
Automation and Mechatronics Engineering,
Material and Chemical Engineering,
Civil and Architecture Engineering,
Biotechnology and Bio Engineering,
Environmental Engineering,
Petroleum and Mining Engineering,
Marine and Agriculture engineering,
Aerospace Engineering.
Surveillance refers to the task of observing a scene, often for lengthy periods in search of particular objects or particular behaviour. This task has many applications, foremost among them is security (monitoring for undesirable behaviour such as theft or vandalism), but increasing numbers of others in areas such as agriculture also exist. Historically, closed circuit TV (CCTV) surveillance has been mundane and labour Intensive, involving personnel scanning multiple screens, but the advent of reasonably priced fast hardware means that automatic surveillance is becoming a realistic task to attempt in real time. Several attempts at this are underway.
Special Plenary Lecture at the International Conference on VIBRATION ENGINEERING AND TECHNOLOGY OF MACHINERY (VETOMAC), Lisbon, Portugal, September 10 - 13, 2018
http://www.conf.pt/index.php/v-speakers
Propagation of uncertainties in complex engineering dynamical systems is receiving increasing attention. When uncertainties are taken into account, the equations of motion of discretised dynamical systems can be expressed by coupled ordinary differential equations with stochastic coefficients. The computational cost for the solution of such a system mainly depends on the number of degrees of freedom and number of random variables. Among various numerical methods developed for such systems, the polynomial chaos based Galerkin projection approach shows significant promise because it is more accurate compared to the classical perturbation based methods and computationally more efficient compared to the Monte Carlo simulation based methods. However, the computational cost increases significantly with the number of random variables and the results tend to become less accurate for a longer length of time. In this talk novel approaches will be discussed to address these issues. Reduced-order Galerkin projection schemes in the frequency domain will be discussed to address the problem of a large number of random variables. Practical examples will be given to illustrate the application of the proposed Galerkin projection techniques.
Video surveillance is becoming more and more important forsocial security, law enforcement, social order,military, and other social problems. In order to manage parking information effectively, this vehicle
detection method is presented. In general, motion detection plays an important role in video surveillance
systems. In this paper, firstly this system uses ViBe method to extract the foreground object, then extracts
HOG features on the performance of the ROI of images. At last this paper presents Support vector machine for vehicle recognition. The results of this test show that, the recognition rate of vehicle’s model in this recognition system is up the industrial application standard.
Video surveillance is becoming more and more important forsocial security, law enforcement, social order,military, and other social problems. In order to manage parking information effectively, this vehicle detection method is presented. In general, motion detection plays an important role in video surveillance systems. In this paper, firstly this system uses ViBe method to extract the foreground object, then extracts HOG features on the performance of the ROI of images. At last this paper presents Support vector machine for vehicle recognition. The results of this test show that, the recognition rate of vehicle’s model in this recognition system is up the industrial application standard.
ideo surveillance is becoming more and more important forsocial security, law enforcement, social
order,military, and other social problems. In order to manage parking information effectively, this vehicle
detection method is presented. In general, motion detection plays an important role in video surveillance
systems. In this paper, firstly this system uses ViBe method to extract the foreground object, then extracts
HOG features on the performance of the ROI of images. At last this paper presents Support vector machine
for vehicle recognition. The results of this test show that, the r
Vehicle Recognition Using VIBE and SVMCSEIJJournal
Video surveillance is becoming more and more important forsocial security, law enforcement, social
order,military, and other social problems. In order to manage parking information effectively, this vehicle
detection method is presented. In general, motion detection plays an important role in video surveillance
systems. In this paper, firstly this system uses ViBe method to extract the foreground object, then extracts
HOG features on the performance of the ROI of images. At last this paper presents Support vector machine
for vehicle recognition. The results of this test show that, the recognition rate of vehicle’s model in this
recognition system is up the industrial application standard.
안녕하세요 딥러닝 논문읽기 모임 입니다! 오늘 소개할 논문은 3D관련 업무를 진행 하시는/ 희망하시는 분들의 필수 논문인 VoxelNET 입니다.
발표자료:https://www.slideshare.net/taeseonryu/mcsemultimodal-contrastive-learning-of-sentence-embeddings
안녕하세요! 딥러닝 논문읽기 모임입니다.
오늘은 자율 주행, 가정용 로봇, 증강/가상 현실과 같은 다양한 응용 분야에서 중요한 문제인 3D 포인트 클라우드에서의 객체 탐지에 대한 획기적인 진전을 소개하고자 합니다. 이를 위해 'VoxelNet'이라는 새로운 3D 탐지 네트워크에 대해 알아보겠습니다.
1. 기존 방법의 한계
기존의 많은 노력은 수동으로 만들어진 특징 표현, 예를 들어 새의 눈 시점 투영 등에 집중해 왔습니다. 하지만 이러한 방법들은 LiDAR 포인트 클라우드와 영역 제안 네트워크(RPN) 사이의 연결을 효과적으로 수행하기 어렵습니다.
2. VoxelNet의 혁신적 접근법
VoxelNet은 3D 포인트 클라우드를 위한 수동 특징 공학의 필요성을 없애고, 특징 추출과 바운딩 박스 예측을 단일 단계, end-to-end 학습 가능한 깊은 네트워크로 통합합니다. VoxelNet은 포인트 클라우드를 균일하게 배치된 3D 복셀로 나누고, 새롭게 도입된 복셀 특징 인코딩(VFE) 레이어를 통해 각 복셀 내의 포인트 그룹을 통합된 특징 표현으로 변환합니다.
3. 효과적인 기하학적 표현 학습
이 방식을 통해 포인트 클라우드는 서술적인 체적 표현으로 인코딩되며, 이는 RPN에 연결되어 탐지를 생성합니다. VoxelNet은 다양한 기하학적 구조를 가진 객체의 효과적인 구별 가능한 표현을 학습합니다.
4. 성능 평가
KITTI 자동차 탐지 벤치마크에서의 실험 결과, VoxelNet은 기존의 LiDAR 기반 3D 탐지 방법들을 큰 차이로 능가했습니다. 또한, LiDAR만을 기반으로 한 보행자와 자전거 탐지에서도 희망적인 결과를 보였습니다.
VoxelNet의 도입은 3D 포인트 클라우드에서의 객체 탐지를 혁신적으로 개선하고 있으며, 이 분야에서의 미래 발전에 중요한 영향을 미칠 것으로 기대됩니다.
오늘 논문 리뷰를 위해 이미지처리 허정원님이 자세한 리뷰를 도와주셨습니다 많은 관심 미리 감사드립니다!
https://youtu.be/yCgsCyoJoMg
A new kind of quantum gates, higher braiding gates, as matrix solutions of the polyadic braid equations (different from the generalized Yang–Baxter equations) is introduced. Such gates lead to another special multiqubit entanglement that can speed up key distribution and accelerate algorithms. Ternary braiding gates acting on three qubit states are studied in detail. We also consider exotic non-invertible gates, which can be related with qubit loss, and define partial identities (which can be orthogonal), partial unitarity, and partially bounded operators (which can be non-invertible). We define two classes of matrices, star and circle ones, such that the magic matrices (connected with the Cartan decomposition) belong to the star class. The general algebraic structure of the introduced classes is described in terms of semigroups, ternary and 5-ary groups and modules. The higher braid group and its representation by the higher braid operators are given. Finally, we show, that for each multiqubit state, there exist higher braiding gates that are not entangling, and the concrete conditions to be non-entangling are given for the obtained binary and ternary gates.
Slides were formed by referring to the text Machine Learning by Tom M Mitchelle (Mc Graw Hill, Indian Edition) and by referring to Video tutorials on NPTEL
FACE RECOGNITION ALGORITHM BASED ON ORIENTATION HISTOGRAM OF HOUGH PEAKSijaia
In this paper we propose a novel face recognition algorithm based on orientation histogram of Hough Transform Peaks. The novelty of the approach lies in utilizing Hough Transform peaks for determining the orientation angles and computing the histogram from it. For extraction of feature vectors first the images are divided into non overlapping blocks of equal size. Then for each of the blocks the orientation histograms are computed. The obtained histograms are combined to form the final feature vector set. Classification is done using k nearest neighbor classifier. The algorithm has been tested on the ORL
database, Yale B Database & the Essex Grimace Database.97% Recognition rates have been obtained for
ORL database, 100% for Yale B and 100% for Essex Grimace database
Data Centers - Striving Within A Narrow Range - Research Report - MCG - May 2...pchutichetpong
M Capital Group (“MCG”) expects to see demand and the changing evolution of supply, facilitated through institutional investment rotation out of offices and into work from home (“WFH”), while the ever-expanding need for data storage as global internet usage expands, with experts predicting 5.3 billion users by 2023. These market factors will be underpinned by technological changes, such as progressing cloud services and edge sites, allowing the industry to see strong expected annual growth of 13% over the next 4 years.
Whilst competitive headwinds remain, represented through the recent second bankruptcy filing of Sungard, which blames “COVID-19 and other macroeconomic trends including delayed customer spending decisions, insourcing and reductions in IT spending, energy inflation and reduction in demand for certain services”, the industry has seen key adjustments, where MCG believes that engineering cost management and technological innovation will be paramount to success.
MCG reports that the more favorable market conditions expected over the next few years, helped by the winding down of pandemic restrictions and a hybrid working environment will be driving market momentum forward. The continuous injection of capital by alternative investment firms, as well as the growing infrastructural investment from cloud service providers and social media companies, whose revenues are expected to grow over 3.6x larger by value in 2026, will likely help propel center provision and innovation. These factors paint a promising picture for the industry players that offset rising input costs and adapt to new technologies.
According to M Capital Group: “Specifically, the long-term cost-saving opportunities available from the rise of remote managing will likely aid value growth for the industry. Through margin optimization and further availability of capital for reinvestment, strong players will maintain their competitive foothold, while weaker players exit the market to balance supply and demand.”
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Data and AI
Discussion on Vector Databases, Unstructured Data and AI
https://www.meetup.com/unstructured-data-meetup-new-york/
This meetup is for people working in unstructured data. Speakers will come present about related topics such as vector databases, LLMs, and managing data at scale. The intended audience of this group includes roles like machine learning engineers, data scientists, data engineers, software engineers, and PMs.This meetup was formerly Milvus Meetup, and is sponsored by Zilliz maintainers of Milvus.
Chatty Kathy - UNC Bootcamp Final Project Presentation - Final Version - 5.23...John Andrews
SlideShare Description for "Chatty Kathy - UNC Bootcamp Final Project Presentation"
Title: Chatty Kathy: Enhancing Physical Activity Among Older Adults
Description:
Discover how Chatty Kathy, an innovative project developed at the UNC Bootcamp, aims to tackle the challenge of low physical activity among older adults. Our AI-driven solution uses peer interaction to boost and sustain exercise levels, significantly improving health outcomes. This presentation covers our problem statement, the rationale behind Chatty Kathy, synthetic data and persona creation, model performance metrics, a visual demonstration of the project, and potential future developments. Join us for an insightful Q&A session to explore the potential of this groundbreaking project.
Project Team: Jay Requarth, Jana Avery, John Andrews, Dr. Dick Davis II, Nee Buntoum, Nam Yeongjin & Mat Nicholas
Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...Subhajit Sahu
Abstract — Levelwise PageRank is an alternative method of PageRank computation which decomposes the input graph into a directed acyclic block-graph of strongly connected components, and processes them in topological order, one level at a time. This enables calculation for ranks in a distributed fashion without per-iteration communication, unlike the standard method where all vertices are processed in each iteration. It however comes with a precondition of the absence of dead ends in the input graph. Here, the native non-distributed performance of Levelwise PageRank was compared against Monolithic PageRank on a CPU as well as a GPU. To ensure a fair comparison, Monolithic PageRank was also performed on a graph where vertices were split by components. Results indicate that Levelwise PageRank is about as fast as Monolithic PageRank on the CPU, but quite a bit slower on the GPU. Slowdown on the GPU is likely caused by a large submission of small workloads, and expected to be non-issue when the computation is performed on massive graphs.
The Building Blocks of QuestDB, a Time Series Databasejavier ramirez
Talk Delivered at Valencia Codes Meetup 2024-06.
Traditionally, databases have treated timestamps just as another data type. However, when performing real-time analytics, timestamps should be first class citizens and we need rich time semantics to get the most out of our data. We also need to deal with ever growing datasets while keeping performant, which is as fun as it sounds.
It is no wonder time-series databases are now more popular than ever before. Join me in this session to learn about the internal architecture and building blocks of QuestDB, an open source time-series database designed for speed. We will also review a history of some of the changes we have gone over the past two years to deal with late and unordered data, non-blocking writes, read-replicas, or faster batch ingestion.
Adjusting OpenMP PageRank : SHORT REPORT / NOTESSubhajit Sahu
For massive graphs that fit in RAM, but not in GPU memory, it is possible to take
advantage of a shared memory system with multiple CPUs, each with multiple cores, to
accelerate pagerank computation. If the NUMA architecture of the system is properly taken
into account with good vertex partitioning, the speedup can be significant. To take steps in
this direction, experiments are conducted to implement pagerank in OpenMP using two
different approaches, uniform and hybrid. The uniform approach runs all primitives required
for pagerank in OpenMP mode (with multiple threads). On the other hand, the hybrid
approach runs certain primitives in sequential mode (i.e., sumAt, multiply).
Techniques to optimize the pagerank algorithm usually fall in two categories. One is to try reducing the work per iteration, and the other is to try reducing the number of iterations. These goals are often at odds with one another. Skipping computation on vertices which have already converged has the potential to save iteration time. Skipping in-identical vertices, with the same in-links, helps reduce duplicate computations and thus could help reduce iteration time. Road networks often have chains which can be short-circuited before pagerank computation to improve performance. Final ranks of chain nodes can be easily calculated. This could reduce both the iteration time, and the number of iterations. If a graph has no dangling nodes, pagerank of each strongly connected component can be computed in topological order. This could help reduce the iteration time, no. of iterations, and also enable multi-iteration concurrency in pagerank computation. The combination of all of the above methods is the STICD algorithm. [sticd] For dynamic graphs, unchanged components whose ranks are unaffected can be skipped altogether.
Adjusting primitives for graph : SHORT REPORT / NOTESSubhajit Sahu
Graph algorithms, like PageRank Compressed Sparse Row (CSR) is an adjacency-list based graph representation that is
Multiply with different modes (map)
1. Performance of sequential execution based vs OpenMP based vector multiply.
2. Comparing various launch configs for CUDA based vector multiply.
Sum with different storage types (reduce)
1. Performance of vector element sum using float vs bfloat16 as the storage type.
Sum with different modes (reduce)
1. Performance of sequential execution based vs OpenMP based vector element sum.
2. Performance of memcpy vs in-place based CUDA based vector element sum.
3. Comparing various launch configs for CUDA based vector element sum (memcpy).
4. Comparing various launch configs for CUDA based vector element sum (in-place).
Sum with in-place strategies of CUDA mode (reduce)
1. Comparing various launch configs for CUDA based vector element sum (in-place).
1. AUTOMATIC DETECTION OF
COMPOUND STRUCTURES
FROM MULTIPLE HIERARCHICAL
SEGMENTATIONS
H¨useyin G¨okhan Akc¸ay
Department of Computer Engineering, Bilkent University, Bilkent, 06800, Ankara
akcay@cs.bilkent.edu.tr
21 Sept. 2016
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 1 / 79
2. Motivation
Large scale global content about the Earth.
Small local details (upto 30 cm resolution).
1.6 terabytes of data by the ESA’s multispectral
high-resolution imaging satellite.
The WorldView-2 satellite collects 975,000 square
kilometers of imagery per day.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 2 / 79
3. Motivation
A challenging problem in remote sensing image mining is
the detection of heterogeneous compound structures such
as different types of residential, industrial, and agricultural
areas.
Compound structures are comprised of spatial
arrangements of simple primitive objects such as buildings,
trees and road segments.
Detection of compound structures is a challenging problem
because
They contain thousands of primitive objects.
They mostly do not have distinctive features.
Primitives can arrange in many different combinations in the
overhead view.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 3 / 79
4. Motivation
Figure: 75 × 75m2 compound structures in WorldView-2 images.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 4 / 79
5. Literature Review
Primitive object detection.
Residential-factory buildings, local roads, vehicles, airplanes
and boats.
Window-based approaches.
Bag-of-words representation.
Enforces artificial boundaries on the image.
Assumes the whole window corresponds to a compound
structure.
Segmentation-based approaches.
The grouping criteria do not involve spatial arrangements.
Graph-based approaches.
Specific arrangements such as alignment and parallelism.
Structural graph matching.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 5 / 79
6. Problem Definition
We propose a generic method for the modeling and
detection of compound structures.
Target structures can involve arrangements of an unknown
number of different types of primitive objects.
The detection task is formulated as the selection of multiple
coherent subsets of candidate regions obtained from
multiple hierarchical segmentations.
To avoid over- or under-segmentation of candidate regions,
we search for the most meaningful regions at different
scales.
We propose a constrained region selection framework
which allows to specify global constraints on the selected
regions.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 6 / 79
8. Compound Structure Model
Primitive Representation
Compound structures are composed of spatial
arrangements of multiple, relatively homogeneous, and
compact primitive objects.
We assume that a compound structure V consists of R
layers of primitive object maps, V = r=1,...,R Vr
.
Figure: Primitive object layers.
Each primitive object vi is represented by an ellipse
vi = (li, si, θi).
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 8 / 79
9. Compound Structure Model
Spatial Arrangement Model
For a given compound structure consisting of N primitive
objects, we construct a neighborhood graph G = (V, E).
V = {v1, . . . , vN} correspond to the individual primitive
objects,
E = r1,r2=1,...,R Er1r2 where Er1r2 denotes the edges
between the vertices at layers Vr1 and Vr2 .
Figure: Neighborhood graph construction for multiple primitive layers.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 9 / 79
10. Compound Structure Model
Spatial Arrangement Model
For each (vi, vj) ∈ E, we compute the following five features:
Distance between the
closest pixels, φ1(vi, vj),
Relative orientation,
φ2(vi, vj),
φ2
φ3
φ1 φ4
Angle between the line joining the centroids of the two
objects and the major axis of vi as the reference object,
φ3(vi, vj),
Distance between the closest antipodal pixels that lie on the
major axes, φ4(vi, vj),
Relative size, φ5(vi, vj).
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 10 / 79
11. Compound Structure Model
Spatial Arrangement Model
We also compute the following four individual features for
each primitive object vi:
Area, φ6(vi),
Eccentricity, φ7(vi),
Solidity, φ8(vi),
Regularity, φ9(vi).
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 11 / 79
12. Compound Structure Model
Spatial Arrangement Model
A one-dimensional marginal histogram Hr1r2
k (Er1r2 ) is
constructed for each pairwise feature φk , k = 1, . . . , 5,
computed over all edges for each pair of layers Vr1 and Vr2 .
Also, a one-dimensional marginal histogram Hr
k (Vr
) is
constructed for each individual feature φk , k = 6, . . . , 8,
computed over all vertices at each layer Vr
.
The concatenation H(V) of all marginal histograms is used
as a non-parametric approximation to the distribution of the
primitive objects in the compound structure.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 12 / 79
13. Compound Structure Model
Spatial Arrangement Model
Figure: Example histograms for the building layers of four different
types of compound structures.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 13 / 79
14. Compound Structure Model
Probabilistic Region Processes
Each primitive object vi (i.e., the ellipse parameters) is
considered a vector-valued random variable.
A compound structure is represented by a set of random
variables that leads to a region process.
The region process is governed by the Gibbs distribution
p(V|β) =
1
Zv
exp βT
H(V) (1)
where β is the parameter vector controlling each histogram
bin, and Zv is the partition function.
A region process is equivalent to a Markov random field
(MRF).
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 14 / 79
15. Learning
Maximum Likelihood Estimation
Suppose that we observe a set of i.i.d. region processes
V = {V1, . . . , VM}.
We can estimate a compound structure model via
maximum likelihood estimation (MLE) of β by maximizing
(β|V) =
M
m=1
log p(Vm|β). (2)
The gradient of the log-likelihood is given by
d (β|V)
dβ
= Ep[H(V)] −
1
M
N
m=1
H(Vm). (3)
We use the stochastic gradient ascent algorithm where the
expectation Ep[H(V)] is approximated by a finite sum of
histograms of samples V(s), s = 1, . . . , S.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 15 / 79
16. Learning
Maximum Likelihood Estimation
Figure: An example iteration for updating β corresponding to the
relative orientation histogram bins.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 16 / 79
17. Learning
Sampling Region Processes
(a) (b) t = 0 (c) t = 50
(d) t = 200 (e) t = 600 (f) t = 1000
Figure: Illustration of the Gibbs sampler for two primitive layers.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 17 / 79
18. Inference and Region Selection
Hierarchical Region Extraction
Given a compound structure model with learned parameter
vector β, we would like to automatically detect all of its
instances in an input image.
The detection problem is posed as the selection of multiple
subgroups of candidate regions coming from multiple
hierarchical segmentations.
Figure: Hierarchical segmentation trees for two primitive layers.
Each selected group of regions constitutes an instance of
the example compound structure in the large image.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 18 / 79
19. Inference and Region Selection
Hierarchical Region Extraction
Given a compound structure model with learned parameter
vector β, we would like to automatically detect all of its
instances in an input image.
The detection problem is posed as the selection of multiple
subgroups of candidate regions coming from multiple
hierarchical segmentations.
Figure: Hierarchical segmentation trees for two primitive layers.
Each selected group of regions constitutes an instance of
the example compound structure in the large image.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 18 / 79
20. Inference and Region Selection
Hierarchical Region Extraction
The first step is the identification of candidate regions for
each layer Vr
by using a hierarchical segmentation
algorithm.
The next step is to connect the potentially related vertices
at all levels to represent the neighbor relationships.
Within-level edges (⊆ Er1r2 , r1 = r2): Voronoi tessellations.
Between-level edges (⊆ Er1r2 , r1 = r2): Ancestor-descendant
relations.
Between-layer edges (⊆ Er1r2 , r1 = r2): Proximity-based
neighbors.
Figure: Hierarchical segmentation trees for two primitive layers.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 19 / 79
21. Inference and Region Selection
Hierarchical Region Extraction
The first step is the identification of candidate regions for
each layer Vr
by using a hierarchical segmentation
algorithm.
The next step is to connect the potentially related vertices
at all levels to represent the neighbor relationships.
Within-level edges (⊆ Er1r2 , r1 = r2): Voronoi tessellations.
Between-level edges (⊆ Er1r2 , r1 = r2): Ancestor-descendant
relations.
Between-layer edges (⊆ Er1r2 , r1 = r2): Proximity-based
neighbors.
Figure: Hierarchical segmentation trees for two primitive layers.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 19 / 79
22. Inference and Region Selection
Hierarchical Region Extraction
Figure: Graph construction for two primitive layers (i.e., building and
pool). The hierarchical candidate regions at three and two levels for
these layers are shown in red and light blue, respectively. The edges
that represent parent-child relationship for both layers are shown.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 20 / 79
23. Inference and Region Selection
Hierarchical Region Extraction
Figure: Graph construction for two primitive layers (i.e., building and
pool). The edges that represent the within- and between-level
neighbor relationship within the same layer are shown.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 21 / 79
24. Inference and Region Selection
Hierarchical Region Extraction
Figure: Graph construction for two primitive layers (i.e., building and
pool). The edges that represent the within- and between-level
neighbor relationship between the layers are shown. For better
visualization of edges, only 20 percent of all between-layer edges are
shown.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 22 / 79
25. Inference and Region Selection Inference without Constraints
Bayesian Formulation
Given a graph G = (V, E), the problem can be formulated
as the selection of a subset V∗
among all regions V as
V∗
= arg max
V ⊆V
p(V |I) = arg max
V ⊆V
p(I|V )p(V ) (4)
where p(I|V ) is the observed spectral data likelihood for
the compound structure in the image, and p(V ) acts as the
spatial prior according to the learned appearance and
arrangement model.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 23 / 79
26. Inference and Region Selection Inference without Constraints
CRF Formulation
We formulate the selection problem in (4) using a
conditional random field (CRF).
Let X = {x1, . . . , xM} where xi ∈ {0, 1}, i = 1, . . . , M, be the
set of indicator variables associated with the vertices V of G
so that xi = 1 implies region vi being selected.
Our CRF formulation defines a posterior distribution as
p(X|I, V) ∝ p(I|X, V)p(X, V)
=
1
Zx
vi ∈V
exp ψc
i + ψs
i xi
(vi ,vj )∈E
exp ψa
ij xixj . (5)
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 24 / 79
27. Inference and Region Selection Inference without Constraints
CRF Formulation
The vertex bias terms ψc
and ψs
representing color and
shape, respectively, and edge weights ψa
representing
arrangement are defined as
ψc
i =
−1
2
(yi − µr
)T
(Σr
)−1
(yi − µr ), ∀vi ∈ Vr
, r = 1, . . . , R
(6)
ψs
i =
9
k=6
βr
k,Ir
k
φk (vi )
, ∀vi ∈ Vr
, r = 1, . . . , R
(7)
ψa
ij =
5
k=1
βr1r2
k,I
r1r2
k
φk (vi ,vj )
, ∀(vi, vj) ∈ E, r1, r2 = 1, . . .
(8)
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 25 / 79
28. Inference and Region Selection Inference without Constraints
CRF Inference
Selecting V∗
in (4) is equivalent to estimating the joint MAP
labels given by
X∗
= arg max
X
p(X|I, V). (9)
Exact inference of the CRF formulation is intractable in
general graphs.
An approximate solution can be obtained by a Markov chain
Monte Carlo sampler.
We developed a sampling algorithm that samples the labels
of many variables at once.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 26 / 79
29. Inference and Region Selection Inference without Constraints
CRF Inference
Figure: Illustration of the primitive sampling procedure.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 27 / 79
30. Inference with Constraints
Our objective is to obtain the maximum probability
estimates of the indicator variables, xi, i = 1, . . . , N
satisfying convex inequality and equality constraints.
We reformulate the problem as quadratic programming
under convex constraints.
The problem in Equation (4) can be rewritten as
V∗
= arg max
V ⊆V
V ⊆Ω
p(V |I) = arg max
V ⊆V
V ⊆Ω
p(I|V )p(V ). (10)
where Ω ∈ RN
is a nonempty polyhedral convex set
determined by a set of constraints.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 28 / 79
31. Inference with Constraints
Quadratic Programming Formulation
For a V ⊆ V, log p(V |β∗
) can be written as follows
log p(V |β) =
5
k=1
R
r1=1
R
r2=1 (vi ,vj )∈Er1r2
βr1r2
k,I
r1r2
k
φk (vi ,vj )
xixj
+
9
k=6
R
r=1 vi ∈Vr
βr
k,Ir
k
φk (vi )
xi − log ZX .
(11)
where ZX is the partition function.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 29 / 79
32. Inference with Constraints
Quadratic Programming Formulation
Let W = 5
k=1
R
r1=1
R
r2=1 Wr1r2
k where each Wr1r2
k is an
N × N affinity matrix.
Each element of this matrix is calculated as
Wr1r2
k (i, j) = −βr1r2
k,I
r1r2
k
φk (vi ,vj )
.
Also, let q = 9
k=6
R
r=1 qr
k where each qr
k is an N × 1
potential vector.
Each element of this vector is calculated as
qr
k (i) = βr
k,Ir
k
φk (vi )
.
The problem can be formulated as
minimize
x
− log p(V |β) =
1
2
XT
WX + qT
X + log ZX
subject to X ∈ Ω,
X ∈ {0, 1}.
(12)
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 30 / 79
33. Inference with Constraints
DC Programming Inference
The problem can be formulated as
minimize
x
− log p(V |β) =
1
2
XT
WX + qT
X + log ZX
subject to X ∈ Ω,
X ∈ {0, 1}.
(13)
First, a linear programming relaxation is applied to the 0 − 1
integer program so that 0 ≤ x ≤ 1.
Since W is not assumed positive semidefinite, the resulting
linearly constrained quadratic problem is not convex.
The objective function can be reformulated as a difference
of two convex functions.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 31 / 79
34. Inference with Constraints
Difference of Convex Programming
Formulation
A Difference of Convex (DC) problem is defined as
P = min f(X) = g(X) − h(X) : X ∈ RN
(14)
where g : RN
→ R and h : RN
→ R are convex functions.
Consider the dual program
D = min f∗
(Y) = h∗
(Y) − g∗
(Y) : Y ∈ RN
(15)
where g∗
is the conjugate function of g.
An iterative primal-dual algorithm constructs two alternating
sequences {X(t)
} and {Y(t)
} such that
g(X(t)) − h(X(t)) and g∗(Y(t)) − h∗(Y(t)) are decreasing,
converging to the optimal solutions, X∗ and Y∗, to the primal
and dual problems, respectively.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 32 / 79
35. Inference with Constraints
DC Programming Inference
Let W = QΛQT
be the eigenvalue decomposition of W, and
Λ+
= diag(Λ+
1 , . . . , Λ+
N ) (respectively, Λ−
= diag(Λ−
1 , . . . , Λ−
N ))
be the positive semidefinite diagonal matrix (respectively,
negative semidefinite diagonal matrix) of Λ.
We rewrite the nonconvex symmetric quadratic objective
function as
1
2
XT
WX + qT
X = g(X) − h(X)
g(X) =
1
2
XT
W+
X + qT
X + χΩ(X)
h(X) = −
1
2
XT
W−
X
(16)
where W+
= QΛ+
QT
, W−
= QΛ−
QT
, and χΩ(X) is the
space enclosed by the constraints X ∈ {0, 1} and Ω.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 33 / 79
36. Inference with Constraints
Experiments
We present detailed results of four different kinds of
experiments:
1 Using a single layer without imposing any constraint.
2 Using multiple layers without imposing any constraint.
3 Using a single layer by imposing constraints.
4 Using multiple layers by imposing constraints.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 34 / 79
37. Experiments Unconstrained Single-layer Experiments
Results-Urban Structures
Figure: 2500 × 4000 pixels Ankara data set.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 35 / 79
38. Experiments Unconstrained Single-layer Experiments
Results-Urban Structures
Table: Detection scenarios for the experiments. Example primitives
used for learning the compound structure model for each scenario are
shown in a different color. The number of polygons and buildings in
the validation data are also given.
Scenario 1 2 3 4 5 6
Example
primitives
# polygons 162 98 48 195 60 16
# buildings 1519 870 1117 1796 771 219
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 36 / 79
39. Experiments Unconstrained Single-layer Experiments
Results-Urban Structures
Figure: Candidate regions hierarchy.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 37 / 79
40. Experiments Unconstrained Single-layer Experiments
Results-Urban Structures
Figure: Marginal probabilities for the first scenario.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 38 / 79
41. Experiments Unconstrained Single-layer Experiments
Results-Urban Structures
Figure: Marginal probabilities for the second scenario.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 39 / 79
42. Experiments Unconstrained Single-layer Experiments
Results-Urban Structures
Figure: Marginal probabilities for the third scenario.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 40 / 79
43. Experiments Unconstrained Single-layer Experiments
Results-Urban Structures
Figure: Marginal probabilities for the fourth scenario.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 41 / 79
44. Experiments Unconstrained Single-layer Experiments
Results-Urban Structures
Figure: Marginal probabilities for the fifth scenario.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 42 / 79
45. Experiments Unconstrained Single-layer Experiments
Results-Urban Structures
*
Figure: Marginal probabilities for the sixth scenario.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 43 / 79
53. Experiments Unconstrained Single-layer Experiments
Results-Orchards
Figure: Example results for the detection of orchards in the subimage
on the left column. The right column shows the corresponding
marginal probabilities of the selected regions (the copper colormap) as
well as the discarded input candidate regions (white).
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 51 / 79
54. Experiments Unconstrained Single-layer Experiments
Results-Orchards
Figure: Example results for the detection of orchards in the subimage
on the left column. The right column shows the corresponding
marginal probabilities of the selected regions (the copper colormap) as
well as the discarded input candidate regions (white).
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 52 / 79
55. Experiments Unconstrained Single-layer Experiments
Results-Orchards
Figure: Example results for the detection of orchards. The left column
shows the marginal probabilities at the end of selection. The right
column shows the thresholded detections overlayed as red.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 53 / 79
56. Experiments Unconstrained Single-layer Experiments
Results-Refugee Camps
Figure: 1102 × 971 pixels Darfur data set.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 54 / 79
57. Experiments Unconstrained Single-layer Experiments
Results-Refugee Camps
Figure: Example results for the detection of refugee camps as rural
structures.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 55 / 79
58. Experiments Unconstrained Multi-layer Experiments
Results-Housing Estates
Figure: 3000 × 8000 pixels Kusadasi data set.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 56 / 79
59. Experiments Unconstrained Multi-layer Experiments
Results-Housing Estates
Figure: (Up) Examples of local details of red building rooftops. (Down)
An example hierarchy.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 57 / 79
60. Experiments Unconstrained Multi-layer Experiments
Results-Housing Estates
The selection algorithm that used only the building layer
could not detect several housing estates.
The idea was to add a pool layer that can provide additional
cues for finding the missed buildings.
The initial model was extended by learning the
arrangements of buildings with respect to pools as well.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 58 / 79
61. Experiments Unconstrained Multi-layer Experiments
Results-Housing Estates
Figure: (Up) Candidate regions. (Down) Selected regions from the
building and pool layers.H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 59 / 79
62. Experiments Unconstrained Multi-layer Experiments
Results-Housing Estates
Table: The number of candidate and detected regions for single and
multi-layer selection scenarios.
Single-layer Multi-layer
Candidates Detected Candidates Detected
Building 67,983 11,173 67,983 11,871
Pool - - 16,276 436
Total 67,983 11,173 84,259 12,307
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 60 / 79
63. Experiments Unconstrained Multi-layer Experiments
Results-Housing Estates
Figure: Samples obtained by the selection procedure ran on single
and multiple layers.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 61 / 79
64. Experiments Unconstrained Multi-layer Experiments
Results-Housing Estates
(a) (b)
Figure: The selected regions using (a) only the building layer. (b)
building and pool layers. Newly detected housing estates that was
missed with single layer selection is enclosed by a red convex hull.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 62 / 79
65. Experiments Unconstrained Multi-layer Experiments
Results-Housing Estates
(a) (b)
Figure: The selected regions using (a) only the building layer. (b)
building and pool layers. Newly detected housing estates that was
missed with single layer selection is enclosed by a red convex hull.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 63 / 79
66. Experiments Unconstrained Multi-layer Experiments
Results-Housing Estates
Figure: Ground view of a missed housing estate with single layer
selection.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 64 / 79
68. Experiments
Problem Definition
Unconstrained selection involved overlapping regions at
different levels of the hierarchy.
To overcome this problem, we require at most one region
should be selected per path where a path corresponds to
the set of vertices from a leaf to the root.
Formally, we select an optimal subset V∗
⊆ V such that
∀a, b ∈ V∗
, a ∈ descendant(b) and b ∈ descendant(a).
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 66 / 79
69. Experiments
Constrained Single-layer Experiments
Let A be a |P| × |V| matrix.
P denotes all the paths from the leaves to the roots of the
input hierarchical forest.
A(i, j) = 1 implies vi ∈ pj ∈ P.
The problem can be reformulated as
minimize
x
1
2
XT
WX + qT
X + log ZX
subject to AX ≤ 1,
0 ≤ X ≤ 1.
(17)
The resulting problem is solved by the DC inference
algorithm.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 67 / 79
70. Experiments
Results-Urban Structures
Table: The number of selected regions for unconstrained and
constrained selection scenarios.
# cand.s 70,644 70,644 70,644 70,644 70,644 22,195
Uncnstr. 3191 1828 3819 3201 2027 1612
Cnstr. 1485 856 2562 1740 811 263
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 68 / 79
71. Experiments
Results-Urban Structures
(a) (b)
Figure: Zoomed detection examples. (a) shows the RGB image for a
300 × 300 sub-scene. (b) shows the hierarchy of candidate regions
(two-level hierarchy from bottom to top).
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 69 / 79
72. Experiments
Results-Urban Structures
(a)
(b)
Figure: Zoomed detection examples. (a) shows the RGB image for a
300 × 300 sub-scene. (b) shows the hierarchy of candidate regions
(six-level hierarchy from bottom to top).
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 70 / 79
73. Experiments
Constrained Multi-layer Experiments
The last set of experiments uses two primitive layers and
enforces geometrical constraints between them.
We search for nearby alike buildings and green areas where
each building group must have a green area in the middle.
We strictly require that the distance between the centroid of
the centroids of a selected group of similar buildings and
the centroid of a selected large green area cannot exceed a
distance threshold δ .
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 71 / 79
74. Experiments
Constrained Multi-layer Experiments
Let V1
and V2
represent the building and green area layers.
The desired set of regions can be selected by
minimize
x
1
2
XT
WX + qT
X + log ZX
subject to AX ≤ 1,
1
k1
vi ∈V1
sh
i xi −
1
k2
vj ∈V2
sh
j xj ≤ δ
1
k1
vi ∈V1
sw
i xi −
1
k2
vj ∈V2
sw
j xj ≤ δ
vi ∈V1
xi = k1
vj ∈V2
xj = k2
0 ≤ X ≤ 1.
(18)
where (sh
i , sw
i ) is the centroid of region vi, k1 and k2 denote
the number of buildings and green areas to be selected.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 72 / 79
75. Experiments Constrained Multi-layer Experiments
Results-Buildings & Green Areas
Figure: 2500 × 4000 pixels Ankara data set.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 73 / 79
76. Experiments Constrained Multi-layer Experiments
Results-Buildings & Green Areas
Figure: Selected regions for the green areas surrounded by buildings.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 74 / 79
77. Experiments Constrained Multi-layer Experiments
Results-Buildings & Green Areas
(a) RGB (b) Building candidates
(c) Green candi-
dates
(d) Selection (e) Overlay
Figure: A zoomed detection example for k1 = 4, k2 = 1.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 75 / 79
78. Experiments Constrained Multi-layer Experiments
Results-Buildings & Green Areas
(a) RGB (b) Building candidates
(c) Green candi-
dates
(d) Selection (e) Overlay
Figure: A zoomed detection example for k1 = 6, k2 = 1.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 76 / 79
79. Experiments Constrained Multi-layer Experiments
Results-Buildings & Green Areas
(a) k1 = 4, fVal = −2.95 (b) k1 = 8, fVal = −3.23
(c) k1 = 6, fVal = −3.16 (d) k1 = 4, fVal = −2.97
Figure: Zoomed detection examples for different values of k1 = 4, 6, 8.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 77 / 79
80. Experiments
Summary & Conclusions
We described a generic method for the modeling and
detection of compound structures that consisted of
arrangements of mostly unknown number of primitives.
The modeling process built an MRF-based contextual
model for the compound structure of interest.
The detection task involved a combinatorial selection
problem where multiple subsets of candidate regions from
multiple hierarchical segmentations were selected.
We also handled hard constraints on the candidate regions.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 78 / 79
81. Experiments
Summary & Conclusions
Experiments using urban, industrial, agricultural and rural
structures showed that the proposed method can provide
good localization of instances of compound structures.
The multi-layered experiments showed that selection of
some objects required the selection of objects in other
layers that had spatial relation with them.
One of the most important bottlenecks in terms of accuracy
was the errors in the input hierarchical segmentations.
Future work includes
Using the detection results for adjusting wrong segmentation
results.
Inferring the primitive objects inside a compound structure.
H. G. Akc¸ay Compound Structure Detection 21 Sept. 2016 79 / 79