The document discusses machine vision and image matching. It begins with definitions of image matching as the process of geometrically positioning two images so their pixels represent the same physical areas. It describes extracting local invariant features from images using methods like Scale Invariant Feature Transform (SIFT) to find correspondences between images for tasks like object recognition and panorama creation despite variations in lighting, viewpoint and scale. SIFT extracts key points from images and represents each with a 128-element feature vector for robust matching between images.
International Journal of Engineering Research and Applications (IJERA) is an open access online peer reviewed international journal that publishes research and review articles in the fields of Computer Science, Neural Networks, Electrical Engineering, Software Engineering, Information Technology, Mechanical Engineering, Chemical Engineering, Plastic Engineering, Food Technology, Textile Engineering, Nano Technology & science, Power Electronics, Electronics & Communication Engineering, Computational mathematics, Image processing, Civil Engineering, Structural Engineering, Environmental Engineering, VLSI Testing & Low Power VLSI Design etc.
International Journal of Engineering Research and Applications (IJERA) is an open access online peer reviewed international journal that publishes research and review articles in the fields of Computer Science, Neural Networks, Electrical Engineering, Software Engineering, Information Technology, Mechanical Engineering, Chemical Engineering, Plastic Engineering, Food Technology, Textile Engineering, Nano Technology & science, Power Electronics, Electronics & Communication Engineering, Computational mathematics, Image processing, Civil Engineering, Structural Engineering, Environmental Engineering, VLSI Testing & Low Power VLSI Design etc.
DIGITAL RESTORATION OF TORN FILMS USING FILTERING T ECHNIQUESAM Publications
The acceptance of digital imaging is motivating many photography enthusiasts to transfer their
photographic archive to digital form. Scans of negatives and positives are preferred to be scanned at high resolution
which makes small cracks and scratches very apparent. These unsightly defects have become an important issue
for consumers. Filtering techniques are used for the restoration process which is fully automatic whereas the existing
systems were semi-automatic or completely manual. The method used for the detection of tear is dilation process and
top-hat transform. Top-hat transform might misinterpret dark brush strokes as cracks. In order to avoid these
unwanted alterations to the original image, brush strokes are separated from the actual cracks using clustering
technique. Tear removal includes order statistics filtering which deals with the reconstruction of missing or
damaged image areas.
Object Detection and tracking in Video SequencesIDES Editor
This paper focuses on key steps in video analysis
i.e. Detection of moving objects of interest and tracking of
such objects from frame to frame. The object shape
representations commonly employed for tracking are first
reviewed and the criterion of feature Selection for tracking is
discussed. Various object detection and tracking approaches
are compared and analyzed.
ABSTRACT : Image registration is an important and fundamental task in image processing used to match two different images. Image registration estimates the parameters of the geometrical transformation model that maps the sensed images back to its reference image. A Feature-Based Approach to automated image-to-image registration is presented. In this paper, various methods are used in different Phases of Image registration. The characteristics of this approach is it combines scale interaction of Discrete wavelets for feature extraction, Scale Invariant Feature Transform (SIFT) for feature matching. Scale-invariant feature transform (or SIFT) is an algorithm in computer vision to detect and describe local features in images. SIFT feature descriptor is invariant to uniform scaling, orientation, and partially invariant to affine distortion and illumination changes.
Implementing Camshift on a Mobile Robot for Person Tracking and Pursuit_ICDMSoma Boubou
This is the slides for a paper presented in ICDM workshop in Vancouver-Canada 2011.
In the paper we describe a Camshift implementation on mobile robotic system for tracking and pursuing a moving person with a monocular camera. Camshift algorithm uses color distribution information to track moving object. It is computationally efficient for working in real-time applications and robust to image noise. It can deal well with illumination changes, shadows and irregular objects motion (linear/non-linear). We compared the Camshift with a HSV color based tracking and our results show that the Camshift method outperformed the HSV color based tracking. Moreover, the former method is much more robust against different illumination conditions.
Paper link:
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=6137446&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D6137446
An evaluation approach for detection of contours with 4 d images a revieweSAT Journals
Abstract Abstract This paper presents a survey of contour detection and the actual use of contour in image processing. Image processing is
an enhanced area in computer science. Contour detection is the part of image processing. Contours are highly depends on quality
of an image. Contour is nothing but the simple boundaries or outlines in an image. Contour detection is nearly related with image
segmentation, classification and recognition of any object in an image. With help of contour detection we can achieve the high
accuracy of the results. Object recognition image retrieval uses the concept of contour detection to achieve the high accuracy in
the results, so it’s an enhanced and popular method in image processing. Active contour model is also one of the main techniques
in contour detection. Active contour is one of the successful models in image processing. This is a modified method of contour
detection. It consists of evolving an image with help of boundaries. Active contour model is also called as snake. Contour
detection plays an important role in recognition.
Keywords: 4D images, Contour Detection, Image Segmentation, Image Classification etc…
A Lecture I gave to an Artificial Intelligence undergraduate class taught by Hien Nguyen, Ph.D. at the University of Wisconsin Whitewater in the fall of 2011
A novel approach to develop a new hybridijitjournal
Trademark Image Retrieval is playing a vital role as a part of CBIR System. Trademark is of great
significance because it carries the status value of any company. To retrieve such a fake or copied
trademark we design a retrieval system which is based on hybrid techniques. It contains a mixture of two
different feature vector which combined together to give a suitable retrieval system. In the proposed system
we extract the corner feature which is applied on an edge pixel image. This feature is used to extract the
relevant image and to more purify the result we apply other feature which is the invariant moment feature.
From the experimental results we conclude that the system is 85 percent efficient.
Object Capturing In A Cluttered Scene By Using Point Feature MatchingIJERA Editor
Capturing means getting or catching. This project contains an algorithm for capturing a specific target based on the points which corresponds between reference and target image. It can capture the objects in-plane rotation and also effective to small amount of out-of plane rotation also. This method of object capturing works best for objects that exhibit in a cluttered texture patterns, which give rise to unique point feature matches. When a part of object is occluded by other objects in the scene, only features of that part are missed. As long as there are enough features detected in the unoccluded part, the object can captured. The local representation is based on the appearance. There is no need to extract geometric primitives (e.g. lines) which are generally hard to detect reliably.
DIGITAL RESTORATION OF TORN FILMS USING FILTERING T ECHNIQUESAM Publications
The acceptance of digital imaging is motivating many photography enthusiasts to transfer their
photographic archive to digital form. Scans of negatives and positives are preferred to be scanned at high resolution
which makes small cracks and scratches very apparent. These unsightly defects have become an important issue
for consumers. Filtering techniques are used for the restoration process which is fully automatic whereas the existing
systems were semi-automatic or completely manual. The method used for the detection of tear is dilation process and
top-hat transform. Top-hat transform might misinterpret dark brush strokes as cracks. In order to avoid these
unwanted alterations to the original image, brush strokes are separated from the actual cracks using clustering
technique. Tear removal includes order statistics filtering which deals with the reconstruction of missing or
damaged image areas.
Object Detection and tracking in Video SequencesIDES Editor
This paper focuses on key steps in video analysis
i.e. Detection of moving objects of interest and tracking of
such objects from frame to frame. The object shape
representations commonly employed for tracking are first
reviewed and the criterion of feature Selection for tracking is
discussed. Various object detection and tracking approaches
are compared and analyzed.
ABSTRACT : Image registration is an important and fundamental task in image processing used to match two different images. Image registration estimates the parameters of the geometrical transformation model that maps the sensed images back to its reference image. A Feature-Based Approach to automated image-to-image registration is presented. In this paper, various methods are used in different Phases of Image registration. The characteristics of this approach is it combines scale interaction of Discrete wavelets for feature extraction, Scale Invariant Feature Transform (SIFT) for feature matching. Scale-invariant feature transform (or SIFT) is an algorithm in computer vision to detect and describe local features in images. SIFT feature descriptor is invariant to uniform scaling, orientation, and partially invariant to affine distortion and illumination changes.
Implementing Camshift on a Mobile Robot for Person Tracking and Pursuit_ICDMSoma Boubou
This is the slides for a paper presented in ICDM workshop in Vancouver-Canada 2011.
In the paper we describe a Camshift implementation on mobile robotic system for tracking and pursuing a moving person with a monocular camera. Camshift algorithm uses color distribution information to track moving object. It is computationally efficient for working in real-time applications and robust to image noise. It can deal well with illumination changes, shadows and irregular objects motion (linear/non-linear). We compared the Camshift with a HSV color based tracking and our results show that the Camshift method outperformed the HSV color based tracking. Moreover, the former method is much more robust against different illumination conditions.
Paper link:
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=6137446&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D6137446
An evaluation approach for detection of contours with 4 d images a revieweSAT Journals
Abstract Abstract This paper presents a survey of contour detection and the actual use of contour in image processing. Image processing is
an enhanced area in computer science. Contour detection is the part of image processing. Contours are highly depends on quality
of an image. Contour is nothing but the simple boundaries or outlines in an image. Contour detection is nearly related with image
segmentation, classification and recognition of any object in an image. With help of contour detection we can achieve the high
accuracy of the results. Object recognition image retrieval uses the concept of contour detection to achieve the high accuracy in
the results, so it’s an enhanced and popular method in image processing. Active contour model is also one of the main techniques
in contour detection. Active contour is one of the successful models in image processing. This is a modified method of contour
detection. It consists of evolving an image with help of boundaries. Active contour model is also called as snake. Contour
detection plays an important role in recognition.
Keywords: 4D images, Contour Detection, Image Segmentation, Image Classification etc…
A Lecture I gave to an Artificial Intelligence undergraduate class taught by Hien Nguyen, Ph.D. at the University of Wisconsin Whitewater in the fall of 2011
A novel approach to develop a new hybridijitjournal
Trademark Image Retrieval is playing a vital role as a part of CBIR System. Trademark is of great
significance because it carries the status value of any company. To retrieve such a fake or copied
trademark we design a retrieval system which is based on hybrid techniques. It contains a mixture of two
different feature vector which combined together to give a suitable retrieval system. In the proposed system
we extract the corner feature which is applied on an edge pixel image. This feature is used to extract the
relevant image and to more purify the result we apply other feature which is the invariant moment feature.
From the experimental results we conclude that the system is 85 percent efficient.
Object Capturing In A Cluttered Scene By Using Point Feature MatchingIJERA Editor
Capturing means getting or catching. This project contains an algorithm for capturing a specific target based on the points which corresponds between reference and target image. It can capture the objects in-plane rotation and also effective to small amount of out-of plane rotation also. This method of object capturing works best for objects that exhibit in a cluttered texture patterns, which give rise to unique point feature matches. When a part of object is occluded by other objects in the scene, only features of that part are missed. As long as there are enough features detected in the unoccluded part, the object can captured. The local representation is based on the appearance. There is no need to extract geometric primitives (e.g. lines) which are generally hard to detect reliably.
Object tracking with SURF: ARM-Based platform ImplementationEditor IJCATR
Several algorithms for object tracking, are developed, but our method is slightly different, it’s about how to adapt and implement such algorithms on mobile platform.
We started our work by studying and analyzing feature matching algorithms, to highlight the most appropriate implementation technique for our case.
In this paper, we propose a technique of implementation of the algorithm SURF (Speeded Up Robust Features), for purposes of recognition and object tracking in real time. This is achieved by the realization of an application on a mobile platform such a Raspberry pi, when we can select an image containing the object to be tracked, in the scene captured by the live camera pi. Our algorithm calculates the SURF descriptor for the two images to detect the similarity therebetween, and then matching between similar objects. In the second level, we extend our algorithm to achieve a tracking in real time, all that must respect raspberry pi performances. So, the first thing is setting up all libraries that the raspberry pi need, then adapt the algorithm with card’s performances. This paper presents experimental results on a set of evaluation images as well as images obtained in real time.
Extraction of Buildings from Satellite ImagesAkanksha Prasad
Buildings are termed as important components for various applications. Building extraction is defined as a sub-problem of Object Recognition. Though, numerous building extraction techniques have been proposed in the literature. But still they often exhibit limited success in the real scenarios. The main purpose of this research is to develop an algorithm which is able to detect and extract buildings from satellite images. In the proposed approach feature-based extraction process is used to extract buildings from satellite images. The overall system is tested and high performance detection is achieved which shows the effectiveness of proposed approach.
Improved Characters Feature Extraction and Matching Algorithm Based on SIFTNooria Sukmaningtyas
According to SIFT algorithm does not have the property of affine invariance, and the high
complexity of time and space, it is difficult to apply to real-time image processing for batch image
sequence, so an improved SIFT feature extraction algorithm was proposed in this paper. Firstly, the MSER
algorithm detected the maximally stable extremely regions instead of the DOG operator detected extreme
point, increasing the stability of the characteristics, and reducing the number of the feature descriptor;
Secondly, the circular feature region is divided into eight fan-shaped sub-region instead of 16 square subregion
of the traditional SIFT, and using Gaussian function weighted gradient information field to construct
the new SIFT features descriptor. Compared with traditional SIFT algorithm, The experimental results
showed that the algorithm not only has translational invariance, scale invariance and rotational invariance,
but also has affine invariance and faster speed that meet the requirements of real-time image processing
applications.
International Journal of Engineering Research and Applications (IJERA) is an open access online peer reviewed international journal that publishes research and review articles in the fields of Computer Science, Neural Networks, Electrical Engineering, Software Engineering, Information Technology, Mechanical Engineering, Chemical Engineering, Plastic Engineering, Food Technology, Textile Engineering, Nano Technology & science, Power Electronics, Electronics & Communication Engineering, Computational mathematics, Image processing, Civil Engineering, Structural Engineering, Environmental Engineering, VLSI Testing & Low Power VLSI Design etc.
Comparison of various Image Registration Techniques with the Proposed Hybrid ...idescitation
Image Registration is termed as the method to
transform different forms of image data into one coordinate
system. Registration is a important part in image processing
which is used for matching the pictures which are obtained at
different time intervals or from various sensors. A broad range
of registration techniques have been developed for the various
types of image data. These techniques are independently
studied for many applications resulting in the large body of
result. Vision is the most advanced of human sensors, so
naturally images play one of the most important roles in
human perception. Image registration is one of the branches
encompassed by the diverse field of digital image processing.
Due to its importance in many application areas as well as
since its nature is complicated; image registration is now the
topic of much recent research. Registration algorithms tend
to compute transformations to set correspondence betweenthe two images. In this paper the survey is done on various
image registration techniques. Also the different techniques
are compared with the proposed system of the projec
"FingerPrint Recognition Using Principle Component Analysis(PCA)”Er. Arpit Sharma
Fingerprint recognition is one of the oldest and most popular biometric technologies and it is used in criminal investigations, civilian, commercial applications, and so on. Fingerprint matching is the process used to determine whether the two sets of fingerprints details come from the same finger or not. This work focuses on feature extraction and minutiae matching stage. There are many matching techniques used for fingerprint recognition systems such as minutiae based matching, pattern based matching, Correlation based matching, and image based matching.
A new method based upon Principal Component Analysis (PCA) for fingerprint enhancement is proposed in this paper. PCA is a useful statistical technique that has found application in fields such as face recognition and image compression, and is a common technique for finding patterns in data of high dimension. In the proposed method image is first decomposed into directional images using decimation free Directional Filter bank DDFB. Then PCA is applied to these directional fingerprint images which gives the PCA filtered images. Which are basically directional images? Then these directional images are reconstructed into one image which is the enhanced one. Simulation results are included illustrating the capability of the proposed method.
SIFT Based Feature Extraction and Matching for Archaeological ArtifactsBIPUL MOHANTO [LION]
An initial try for any unknown artifacts match with the existing artifacts to give us matching result. Now this one used for images, extension is under develop for 3D scanned object.
Image morphing provides the tool to generate the flexible and powerful visual effect. Morphing depicts the transformation of one image into another image. The process of image morphing starts with the feature specification phase and then proceeds to warp generation phase, followed by the transition control phase. This paper surveys the various techniques available for all three stages of image morphing.
Hybrid optimization of pumped hydro system and solar- Engr. Abdul-Azeez.pdffxintegritypublishin
Advancements in technology unveil a myriad of electrical and electronic breakthroughs geared towards efficiently harnessing limited resources to meet human energy demands. The optimization of hybrid solar PV panels and pumped hydro energy supply systems plays a pivotal role in utilizing natural resources effectively. This initiative not only benefits humanity but also fosters environmental sustainability. The study investigated the design optimization of these hybrid systems, focusing on understanding solar radiation patterns, identifying geographical influences on solar radiation, formulating a mathematical model for system optimization, and determining the optimal configuration of PV panels and pumped hydro storage. Through a comparative analysis approach and eight weeks of data collection, the study addressed key research questions related to solar radiation patterns and optimal system design. The findings highlighted regions with heightened solar radiation levels, showcasing substantial potential for power generation and emphasizing the system's efficiency. Optimizing system design significantly boosted power generation, promoted renewable energy utilization, and enhanced energy storage capacity. The study underscored the benefits of optimizing hybrid solar PV panels and pumped hydro energy supply systems for sustainable energy usage. Optimizing the design of solar PV panels and pumped hydro energy supply systems as examined across diverse climatic conditions in a developing country, not only enhances power generation but also improves the integration of renewable energy sources and boosts energy storage capacities, particularly beneficial for less economically prosperous regions. Additionally, the study provides valuable insights for advancing energy research in economically viable areas. Recommendations included conducting site-specific assessments, utilizing advanced modeling tools, implementing regular maintenance protocols, and enhancing communication among system components.
Event Management System Vb Net Project Report.pdfKamal Acharya
In present era, the scopes of information technology growing with a very fast .We do not see any are untouched from this industry. The scope of information technology has become wider includes: Business and industry. Household Business, Communication, Education, Entertainment, Science, Medicine, Engineering, Distance Learning, Weather Forecasting. Carrier Searching and so on.
My project named “Event Management System” is software that store and maintained all events coordinated in college. It also helpful to print related reports. My project will help to record the events coordinated by faculties with their Name, Event subject, date & details in an efficient & effective ways.
In my system we have to make a system by which a user can record all events coordinated by a particular faculty. In our proposed system some more featured are added which differs it from the existing system such as security.
Automobile Management System Project Report.pdfKamal Acharya
The proposed project is developed to manage the automobile in the automobile dealer company. The main module in this project is login, automobile management, customer management, sales, complaints and reports. The first module is the login. The automobile showroom owner should login to the project for usage. The username and password are verified and if it is correct, next form opens. If the username and password are not correct, it shows the error message.
When a customer search for a automobile, if the automobile is available, they will be taken to a page that shows the details of the automobile including automobile name, automobile ID, quantity, price etc. “Automobile Management System” is useful for maintaining automobiles, customers effectively and hence helps for establishing good relation between customer and automobile organization. It contains various customized modules for effectively maintaining automobiles and stock information accurately and safely.
When the automobile is sold to the customer, stock will be reduced automatically. When a new purchase is made, stock will be increased automatically. While selecting automobiles for sale, the proposed software will automatically check for total number of available stock of that particular item, if the total stock of that particular item is less than 5, software will notify the user to purchase the particular item.
Also when the user tries to sale items which are not in stock, the system will prompt the user that the stock is not enough. Customers of this system can search for a automobile; can purchase a automobile easily by selecting fast. On the other hand the stock of automobiles can be maintained perfectly by the automobile shop manager overcoming the drawbacks of existing system.
Student information management system project report ii.pdfKamal Acharya
Our project explains about the student management. This project mainly explains the various actions related to student details. This project shows some ease in adding, editing and deleting the student details. It also provides a less time consuming process for viewing, adding, editing and deleting the marks of the students.
Courier management system project report.pdfKamal Acharya
It is now-a-days very important for the people to send or receive articles like imported furniture, electronic items, gifts, business goods and the like. People depend vastly on different transport systems which mostly use the manual way of receiving and delivering the articles. There is no way to track the articles till they are received and there is no way to let the customer know what happened in transit, once he booked some articles. In such a situation, we need a system which completely computerizes the cargo activities including time to time tracking of the articles sent. This need is fulfilled by Courier Management System software which is online software for the cargo management people that enables them to receive the goods from a source and send them to a required destination and track their status from time to time.
Saudi Arabia stands as a titan in the global energy landscape, renowned for its abundant oil and gas resources. It's the largest exporter of petroleum and holds some of the world's most significant reserves. Let's delve into the top 10 oil and gas projects shaping Saudi Arabia's energy future in 2024.
Quality defects in TMT Bars, Possible causes and Potential Solutions.PrashantGoswami42
Maintaining high-quality standards in the production of TMT bars is crucial for ensuring structural integrity in construction. Addressing common defects through careful monitoring, standardized processes, and advanced technology can significantly improve the quality of TMT bars. Continuous training and adherence to quality control measures will also play a pivotal role in minimizing these defects.
Welcome to WIPAC Monthly the magazine brought to you by the LinkedIn Group Water Industry Process Automation & Control.
In this month's edition, along with this month's industry news to celebrate the 13 years since the group was created we have articles including
A case study of the used of Advanced Process Control at the Wastewater Treatment works at Lleida in Spain
A look back on an article on smart wastewater networks in order to see how the industry has measured up in the interim around the adoption of Digital Transformation in the Water Industry.
CFD Simulation of By-pass Flow in a HRSG module by R&R Consult.pptxR&R Consult
CFD analysis is incredibly effective at solving mysteries and improving the performance of complex systems!
Here's a great example: At a large natural gas-fired power plant, where they use waste heat to generate steam and energy, they were puzzled that their boiler wasn't producing as much steam as expected.
R&R and Tetra Engineering Group Inc. were asked to solve the issue with reduced steam production.
An inspection had shown that a significant amount of hot flue gas was bypassing the boiler tubes, where the heat was supposed to be transferred.
R&R Consult conducted a CFD analysis, which revealed that 6.3% of the flue gas was bypassing the boiler tubes without transferring heat. The analysis also showed that the flue gas was instead being directed along the sides of the boiler and between the modules that were supposed to capture the heat. This was the cause of the reduced performance.
Based on our results, Tetra Engineering installed covering plates to reduce the bypass flow. This improved the boiler's performance and increased electricity production.
It is always satisfying when we can help solve complex challenges like this. Do your systems also need a check-up or optimization? Give us a call!
Work done in cooperation with James Malloy and David Moelling from Tetra Engineering.
More examples of our work https://www.r-r-consult.dk/en/cases-en/
Industrial Training at Shahjalal Fertilizer Company Limited (SFCL)MdTanvirMahtab2
This presentation is about the working procedure of Shahjalal Fertilizer Company Limited (SFCL). A Govt. owned Company of Bangladesh Chemical Industries Corporation under Ministry of Industries.
Sachpazis:Terzaghi Bearing Capacity Estimation in simple terms with Calculati...Dr.Costas Sachpazis
Terzaghi's soil bearing capacity theory, developed by Karl Terzaghi, is a fundamental principle in geotechnical engineering used to determine the bearing capacity of shallow foundations. This theory provides a method to calculate the ultimate bearing capacity of soil, which is the maximum load per unit area that the soil can support without undergoing shear failure. The Calculation HTML Code included.
2. Machine Vision
LEARNING OUTCOMES
a. Peserta diharapkan memahami definisi dari proses matching serta menerapkannya pada
permasalahan sehari-hari.
b. Peserta dapat memahami fitur lokal yang bersifat invariant dan penerapannya pada proses
matching.
c. Peserta dapat memahami dan mengimplementasikan metode ekstraksi fitur menggunakan
Scale Infariant Feature Transform (SIFT)
OUTLINE MATERI (Sub-Topic):
1. Definition
2. Invariant Local Features
3. Scale Infariant Feature Transform (SIFT)
4. Feature Matching
3. Machine Vision
ISI MATERI
Image Matching
Image matching didefinisikan sebagai sebuah proses untuk memposisikan dua citra
secara geometris agar piksel-piksel pada kedua citra tersebut merepresentasikan area fisik
yang sama dari obyek/pemandangan pada citra tersebut. Proses matching dilakukan dengan
menerapkan transformasi geometrik seperti translasi, rotasi, skala, dan lain-lain sedemikian
sehingga kesamaan antara dua citra yang dicocokkan menjadi maksimal. Obyek pada dunia
nyata didefinisikan pada tiga dimensi, sehingga proses pencocokkan akan lebih sulit
dilakukan apabila sebuah obyek dilihat dari beberapa sudut pandang berbeda dan memiliki
kondisi pencahayaan yang sangat bervariasi. Sebagai contoh akan dilakukan penggabungan
beberapa buah citra untuk membentuk panorama pegunungan, dimana citra-citra tersebut
diambil pada waktu yang berbeda dengan lokasi pengambilan yang berbeda. Gambar berikut
mengilustrasikan proses pembentukan panorama pegunungan dari 5 citra berbeda.
Pengambilan gambar pada waktu yang berbeda akan menghasilkan pencahayaan yang
berbeda, penggabungan dari citra tersebut membutuhkan penyesuaian tingkat pencahayaan.
Meskipun obyek yang diambil gambarnya adalah sama, namun lokasi pengambilan gambar
yang berbeda akan menghasilkan obyek yang tampak berbeda pada citra yang dihasilkan.
Perlu dilakukan koreksi geometri agar kedua gambar tersebut dapat digabungkan.
Image matching merupakan salah satu aspek mendasar dari beberapa permasalahan di
bidang computer vision, seperti pengenalan obyek, content-based image retrieval, stereo
correspondence¸ deteksi gerakan, klasifikasi tekstur dan video data mining. Hingga saat ini
image matching merupakan permasalahan yang kompleks karena beberapa hal, seperti obyek
4. Machine Vision
yang sebagian terhalang oleh obyek lain (partial occlusion), perbedaan kondisi saat
pengambilan gambar (sudut pandang maupun pencahayaan).
Berikut tahapan dari image matching:
1. Mencari area yang dapat dijadikan sebagai kunci dalam menggabungkan dua citra, area
tersebut harus ada di kedua citra dan memiliki ciri yang khusus sehingga dapat dibedakan
dari area lainnya. Dalam contoh pembuatan panorama pegunungan, area yang dipilih
dapat berupa puncak gunung, lembah, atau obyek lainnya.
2. Mendapatkan deskripsi dari area yang dipilih dengan cara menghitung fitur lokal yang
bersifat invariant (tidak dipengaruhi oleh operasi transformasi geometrik seperti translasi,
rotasi, maupun skala).
3. Mendefinisikan korespondensi pada dua citra yang akan digabungkan berdasarkan
kesamaan fiturnya.
4. Ulangi langkah 1 sampai 3 untuk mencari area-area lain yang memiliki korespondensi
pada dua citra tersebut.
Gambar berikut memperlihatkan pemilihan area-area yang dapat dijadikan kunci untuk
menggabungkan dua citra truk mainan. Dari 5 total area yang dipilih, terdapat 3 area yang
saling berkorespondensi pada kedua citra tersebut, yaitu yang ditandai dengan bujur sangkar
berwarna biru.
5. Machine Vision
ISI MATERI
Invariant Local Features
Fitur lokal (local feature) adalah sebuah pola pada citra yang memiliki karakteristik
berbeda dari area di sekitarnya. Biasanya tampak pada area yang ditandai dengan adanya
perubahan karakteristik tertentu, seperti intentitas keabuan, warna, atau tekstur. Fitur lokal
dapat berupa titik, edge (sisi), atau area kecil pada citra. Untuk mendapatkan deskripsi dari
fitur lokal, beberapa pengukuran dilakukan pada area yang ditentukan. Proses pengukuran
tersebut disebut juga sebagai ekstraksi fitur. Algoritma ekstraksi fitur yang baik akan
menghasilkan pengukuran yang serupa meskipun kondisi citra bervariasi, atau disebut juga
sebagai invariant. Fitur lokal yang bersifat invariant (local invariant features) tidak
dipengaruhi oleh variasi geometrik seperti translasi, rotasi, maupun skala, serta variasi
fotometrik seperti tingkat kecerahan warna, exposure, dan lain-lain.
Misalkan akan dibandingkan dua buah citra I1 dan I2. Kedua citra tersebut
merepresentasikan obyek yang sama dengan beberapa variasi tertentu. Citra I2 bisa saja
merupakan hasil transformasi dari I1. Fitur yang bersifat invariant akan menghasilkan
pengukuran yang serupa dari kedua citra tersebut. Sifat tersebut disebut juga sebagai
transformational invariance. Mayoritas algoritma feature extraction dirancang untuk bersifat
invariant terhadap transformasi dua dimensi, seperti translasi, rotasi, maupun skala. Beberapa
algortima lain juga bersifat invariant terhadap transformasi tiga dimensi seperti misalnya
algoritma SIFT (Scale Invariant Feature Transform) yang tidak dipengaruhi oleh operasi
rotasi hingga 60 derajat.
Fitur yang bersifat invariant biasanya menangkap informasi tidak hanya di sebuah titik
saja, namun juga di area sekeliling titik tersebut. Bentuk paling sederhana dari area dapat
berupa bujur sangkar yang berukuran n piksel. Untuk mengekstraksi fitur menggunakan
fungsi (misalkan f) yang bersifat scale invariant, dapat dilakukan dengan mencari titik yang
memberikan local maximum dari f pada skala yang berbeda. Salah satu fungsi yang umum
digunakan untuk mendefinisikan f adalah Laplacian atau selisih antara dua citra yang dikenai
operasi Gaussian filtering (difference of Gaussian) menggunakan parameter σ berbeda.
Gambar berikut mengilustrasikan ide ini, f yang merupakan fungsi difference of Gaussian
6. Machine Vision
akan menghasilkan local maximum pada posisi yang sama (ditandai dengan huruf ‘x’
berwarna kuning) meskipun digunakan skala atau parameter σ berbeda (𝜎1, 𝜎�, 𝜎�):
𝜎1 𝜎� 𝜎�
Sifat rotation invariant dapat diperoleh dengan cara memutar area citra yang akan dipilih
sebagai kunci dengan sudut tertentu sebelum dilakukan ekstraksi fitur. Besarnya sudut
perputaran dapat diperkirakan dengan menghitung orientasi dominannya. Salah satu
pendekatan yang dapat digunakan untuk mengkalkulasi orientasi dominan adalah menghitung
vektor eigen pada Harris corner detector (materi ini sudah dibahas pada materi kuliah sesi ke
6). Ilustrasi berikut menggambarkan bagaimana pendekatan ini bekerja. Kotak berwarna
putih pada citra merepresentasikan orientasi dominan dari area di sekitar puncak gunung.
Orientasi dominan dari area yang dibatasi oleh kotak tersebut dinyatakan oleh sudut
kemiringan dari kotak tersebut.
7. Machine Vision
ISI MATERI
Scale Infariant Feature Transform (SIFT)
Scale Invariant Feature Transform (SIFT) adalah sebuah metode ektraksi fitur yang
dikembangkan oleh David Lowe (1999, 2004). Metode ini banyak diterapkan pada bidang
computer vision terutama dalam point matching diantara dua citra yang diambil dari sudut
pandang berbeda dan pengenalan obyek. Fitur SIFT bersifat invariant terhadap operasi
translasi, rotasi dan skala, juga tahan terhadap transformasi perspektif dan perbedaan
pencahayaan. Hasil eksperimen membuktikan bahwa SIFT sangat cocok diterapkan untuk
kebutuhan image matching dan pengenalan obyek pada kondisi dunia nyata. Pada formulasi
yang diusulkan oleh David Lowe, SIFT terdiri dari metode untuk mendeteksi interenst point
dari citra graylevel dan perhitungan statistik gradien lokal dari intensitas keabuan citra untuk
memberikan deskripsi struktur citra secara lokal di sekitar interest points.
Berikut karakteristik dari SIFT descriptor:
1. Tahan terhadap perubahan sudut pandang
2. Tahan terhadap rotasi hingga 60 derajat
3. Tahan terhadap perubahan tingkat pencahayaan, dalam beberapa kasus citra yang diambil
pada siang hari dan malam hari.
4. Cepat dan efisien
5. Kode sumber banyak tersedia di intenet:
http://people.csail.mit.edu/albert/ladypack/wiki/index.php/Known_implementations_of_S
IFT
Pada gambar berikut diperlihatkan bahwa SIFT berhasil mendeteksi interest points pada
bangunan gedung dengan baik meskipun terhalang oleh obyek manusia dan terdapat
perbedaan tingkat pencahayaan yang besar.
8. Machine Vision
Ide dasar dari SIFT dapat dijelaskan dalam empat tahapan berikut:
1. Deteksi interest points
2. Tentukan window berukuran 16 × 16 untuk setiap interest points
3. Hitung orientasi (edge orientation) dari masing-masing piksel
4. Eliminasi edge-edge yang memiliki gradient magnitude dibawah nilai threshold (weak
edges)
5. Buat histogram dari orientasi edge yang memiliki gradient magnitude diatas nilai
threshold (surviving edges)
Gambar berikut mengilustrasikan tahapan diatas.
Ekstraksi fitur menggunakan SIFT akan mentranformasikan citra input menjadi koleksi
local feature vector, dimana setiap feature vector tersebut bersifat invariant terhadap operasi
translasi, rotasi, maupun skala. Algoritma SIFT menerapkan empat tahap filter dalam
mengekstrak fitur tersebut:
9. Machine Vision
1. Scale-space Extrema Detection
Tahap ini bertujuan untuk memperoleh lokasi dan skala yang dapat diidentifikasi dari
sudut pandang berbeda dari sebuah obyek yang sama. Hal ini dapat dilakukan dengan
menerapkan fungsi “scale space”:
L(x,y,σ) = G(x,y,σ) * I(x,y)
dimana * adalah operator konvolusi, G(x,y,σ) adalah variable-scale Gaussian, dan I(x,y)
adalah citra input.
Terdapat beberapa pendekatan yang dapat digunakan untuk mendeteksi lokasi keypoints
pada scale-space, salah satunya adalah Difference of Gaussian (DoG), yang mendeteksi
scale-space extrema, D(x,y,σ) dengan mengkalkulasi selisih antara dua citra, dimana salah
satunya k × lebih besar dari yang lainnya. D(x,y,σ) dapat dihitung berdasarkan rumus:
D(x,y,σ) = L(x,y,kσ) - L(x,y,σ)
Untuk mendeteksi lokal maksimum dan lokal minimum dari D(x,y,σ), setiap titik
dibandingkan dengan 8 piksel tetangga pada skala yang sama dan 9 piksel pada satu skala
diatas dan dibawahnya. Jika nilai ditemukan nilainya maksimum/minimum terhadap
semua tetangga tersebut, maka titik tersebut merupakan titik maksimum/minimum.
2. Keypoint Localization
Tahap ini bertujuan untuk mengeliminasi keypoint yang memiliki kontras rendah atau
letak titik sulit ditemukan pada sebuah edge (poorly localised on an edge). Hal ini dapat
dilakukan dengan cara menghitung Laplacian, untuk setiap keypoint pada tahap 1, lokasi
dari extremum z dapat dihitung berdasarkan:
𝑧 = −
���−1
�𝑥�
��
�𝑥
Apabila z dibawah nilai threhsold maka titik tersebut dieliminasi. Persamaan tersebut
akan mengeliminasi keypoint yang memiliki kontras rendah. Titik-titik yang bersifat
“poorly localised on an edge” dapat dieliminasi berdasarkan rasio dari vektor eigen
terbesar terhadap vektor eigen terkecil dari matriks Hessian yang berukuran 2x2 pada
lokasi titik tersebut berada.
10. Machine Vision
3. Orientation Assignment
Tahap ini akan bertujuan untuk menentukan orientasi dari keypoint berdasarkan sifat lokal
dari citra. Deskriptor dari keypoint, dapat direpresentasikan relatif terhadap orientasinya
dengan cara melakukan normalisasi sudut (rotasi sebesar sudut orientasi). Proses
normalisasi akan menghasilkan fitur yang bersifat invariant terhadap operasi rotasi.
Orientasi dari keypoint dapat diperoleh dengan cara berikut:
- Hitung gradient magnitude, m:
𝑚(𝑥, 𝑦) = �(𝐿(𝑥 + 1, 𝑦) − 𝐿(𝑥 − 1, 𝑦))� + (𝐿(𝑥, 𝑦 + 1) − 𝐿(𝑥, 𝑦 − 1))�
- Hitung orientasi, θ:
𝜃(𝑥, 𝑦) = tan−1
�
𝐿(𝑥, 𝑦 + 1) − 𝐿(𝑥, 𝑦 − 1)
𝐿(𝑥 + 1, 𝑦) − 𝐿(𝑥 − 1, 𝑦)
�
- Hitung histogram orientasi dari gradient orientation dari titik-titik sampel
- Tentukan lokasi puncak tertinggi pada histogram kemudian gunakan puncak tersebut
dan puncak-puncak lainnya yang memiliki tinggi di atas 80% dari puncak tertinggi
untuk membentuk keypoint pada orientasi θ.
- Cocokkan (fit) parabola pada 3 nilai terdekat ke puncak tertinggi untuk
menginterpolasi posisi puncak.
4. Orientation Assignment
Informasi gradien lokal yang diperoleh pada tahap sebelumnya, juga digunakan untuk
membentuk keypoint descriptor. Informasi gradien diputar sebesar sudut orientasi θ dan
selanjutnya diberi bobot 1.5 × skala keypoint. Data ini selanjutnya digunakan untuk
membentuk histogram dari setiap window yang berpusat pada keypoint. Keypoint
descriptor biasanya terdiri dari 16 histogram yang tersusun dalam kisi berukuran 4 × 4,
dimana masing-masing tersusun atas 8 bin orientasi yang mewakili setiap arah mata angin
(Utara, Timur Laut, Timur, Tenggara, Selatan, Barat Daya, Barat, Barat Laut). Sehingga
menghasilkan feature vector yang mengandung 128 elemen. Vektor yang dihasilkan
disebut juga sebagai SIFT key.
11. Machine Vision
ISI MATERI
Feature Matching
Jika diketahui sebuah fitur di citra I1, bagaimana menemukan lokasi yang paling cocok
berdasarkan fitur tersebut pada citra I2. Hal tersebut dapat dilakukan dengan mendefinisikan
sebuah fungsi yang mengukur perbedaan atau jarak diantara fitur pada masing-masing citra.
Selanjutnya fungsi tersebut diterapkan pada citra I2, cari lokasi yang menghasilkan perbedaan
yang paling minimum. Bagaimana mendefinisikan fungsi yang dapat mengukur
perbedaan/jarak antara fitur f1 dan f2? Salah satu pendekatan yang umum digunakan adalah
mengukur jarak menggunakan Euclidean distance:
𝑑(𝑢, 𝑣) = ��(𝑢� − 𝑣�)�
�
�
1/�
Proses matching dapat dilakukan dengan salah satu dari strategi berikut:
a. Mengembalikan semua titik yang memiliki jarak d yang lebih kecil dari threshold.
b. Nearest neighbor: titik yang memiliki jarak d terkecil.
c. Nearest neighbor distance ratio:
NNDR = d1/d2
dimana d1, d2 menyatakan jarak ke titik terdekat dan titik terdekat kedua. Apabila nilai
NNDR kecil, maka titik tersebut diasumsikan match.
12. Machine Vision
SIMPULAN
1. Proses image matching merupakan permasalahan yang cukup kompleks karena citra yang
dibandingkan seringkali memiliki perbedaan dalam hal sudut pandang, lokasi, orientasi,
skala, maupun tingkat pencahayaan.
2. Fitur lokal yang bersifat invariant (local invariant features) tidak dipengaruhi oleh variasi
geometrik seperti translasi, rotasi, maupun skala, serta variasi fotometrik seperti tingkat
kecerahan warna, atau exposure sangat penting dalam proses image matching.
3. SIFT merupakan salah satu metode ekstraksi fitur yang memiliki karakteristik: tahan
terhadap perubahan sudut pandang, tahan terhadap rotasi hingga 60 derajat, tahan
terhadap perubahan tingkat pencahayaan, serta cepat dan efisien.
13. Machine Vision
DAFTAR PUSTAKA
1. Forsyth. (2011). Computer Vision a Modern Approach (2nd
Edition). Prentice Hall.
New Jersey. ISBN-10: 013608592X. ISBN-13: 978-0136085928.
2. M. Brown, R. Szeliski, S. Winder, Multi-image matching using multi-scale oriented
patches, CVPR 2005, pp. 510-517 vol. 1.
3. Multi-Image Matching using Multi-Scale Oriented Patches,
https://www.microsoft.com/en-us/research/wp-content/uploads/2004/12/tr-2004-
133.pdf
4. D. G. Lowe, “Distinctive image features from scale-invariant keypoints,” Int. Journal
of Compututer Vision, vol. 60, no. 2, pp. 91–110, Nov. 2004.
5. Known implementations of SIFT,
http://people.csail.mit.edu/albert/ladypack/wiki/index.php?title=Known_implementati
ons_of_SIFT
6. SIFT on OpenCV,
http://docs.opencv.org/2.4/modules/nonfree/doc/feature_detection.html
7. G. P. Kusuma, A. Szabo, L. Yiqun and J. A. Lee, Appearance-based object
recognition using weighted longest increasing subsequence, ICPR 2012, pp. 3668-
3671.
8. K. D. Harjono and G. P. Kusuma, Object instance recognition using best increasing
subsequence, KICSS 2016, Yogyakarta, pp. 1-5.