SlideShare a Scribd company logo
Vote Aggregation Techniques in
the Geo-Wiki Crowdsourcing Game:
A Case Study
Michael Khachay, Oleg Nurmukhametov
Artem Baklanov
Krasovskii Institute of Mathematics and Mechanics
Postdoctoral Research Scholar, ASA, IIASA
Krasovskii Institute of Mathematics and Mechanics
Steffen Fritz, Carl Salk, Linda See, and Dmitry
Shchepashchenko
IIASA
Young  Scientists  Summer  Program
Since  1977,  IIASA’s  annual  3-­month  Young  Scientists  
Summer  Program  (YSSP)  offers  opportunities  to  
talented  young  researchers.
Crowdsourcing = Crowd + Outsourcing
Crowdsourcing is a new approach to perform
tasks, when a group of distributed worldwide
people in total can substitute an expert.
Example:
GEO-wiki project, ESM, IIASA
Geo-Wiki: Cropland Capture Game
5 millions votes
170 000 satellite images
3000 volunteers
6 months
Goal: land cover map
www.geo-wiki.org d
Preprocessing of images
Blur detection
Duplicate detection
Benchmarking
of algorithms for vote
aggregation
Outline of presentation
Noise reduction Prediction
Preprocessing of images
Duplicate detection
Blur detection
Benchmarking
of algorithms for vote
aggregation
Outline of presentation
Noise reduction Prediction
Blur detection algorithm [H Tang, 2012].
Input: image
Output: coefficient of blur in [0, 100]
Detection of blurry images
100%0%
High quality images
Probability
0%
10%
30%
Blurry images
80% 90% 95%
98% 99%
The  table  and  figure  are  from
Tong,  Hanghang,  et  al.  "Blur  detection  for  digital  
images  using  wavelet  transform." Multimedia  
and  Expo,  2004.  ICME'04.  2004  IEEE  
International  Conference  on.  Vol.  1.  IEEE,  2004.
Blur detection algorithm
A  key  idea
• A= a  number  of  Roof-­Structure  and  
Gstep-­Structure  edge  points  that  have  lost  
their  sharpness.
• B=  a  number  of  Roof-­Structure  and  
Gstep-­Structure  edge  points.  
• Blur  coefficient  =  100*A/B
Detection of blurry images
Results
Current campaign:
Noise reduction in dataset.
+ for future campaigns:
Less workload, more joy.
Blur coefficient Number of images
0% (Excellent images) 74 000 (38%)
80% - 100% (Extremely bad images) 3 200 (2%)
Extremely bad images were removed
Detection of duplicates
Web Link
Image unique
identifier Link to image
1 http://cg.tuwien.ac.at/~sturn/crop/img_78.1875_52.9958_1000.jpg
2 http://cg.tuwien.ac.at/~sturn/crop/img_78.1875_52.9958_500.jpg
3 http://cg.tuwien.ac.at/~sturn/crop/img_78.1875_52.9958_300.jpg
… …
Binary file Image content
Detection of duplicates
Find the 10 differences
Detection of duplicates
The pixel-by-pixel differences
Links: different
Binary files: different
Image content: similar
pHash (perceptual hash) [Zauner, 2010].
Detection of duplicates
pHash.  High  level  description
1. Resize  to  32x32  and  grayscale;;
2. Perform  DCT-­2  transformation  and  keep  
top-­left  8x8;;
3. Calculate  the  average  value  of  
alternating  components;;
4. Set  the  64  hash  bits  to  0  (or  1)  on  
whether   DCT  values  more  (or  less)  the  
average  value.
Detection of copies
Method Copies count Time to completion
Binary data comparison (MD5) 2 700 Fast (~10 min)
Image content comparison (pHash) 10 000 (6%) Slow(~7 hours)
Results
Current campaign:
Increase of statistically significance;
Reduction of dimensionality.
+ for future campaigns:
Less workload, more joy.
Votes  were  merged.
Preprocessing of images
Duplicate detection
Blur detection
Benchmarking
of algorithms for vote
aggregation
Outline of presentation
Noise reduction Prediction
A  part  of  the  dataset was  annotated  by  an  
expert  after  the  campaign  took  place:
• 854 images;;
• 1813 volunteers;;
• 16  940 votes.  
We    sample  two  subsets  for  training  and  testing  
70/30 ratio
Expert dataset
Baseline  algorithms  
We  use  SVD  to  reduce  dimensionality  and  
10-­fold  cross-­validation  to  fit  parameters
A  publicly  available  code  implements  KOS and  EM algorithm;;  
both  are  implemented  in  conjunction  with  reputation  
algorithm  (also  called  Hard  penalty).  
https://github.com/ashwin90/Penalty-­based-­clustering  
[EM]  Dawid,  A.P.,  Skene,  A.M.:  Maximum  likelihood  estimation  of  
observer  error-­rates  using  the  em algorithm.  Applied  statistics  pp.  
20–28  (1979)
[Hard  penalty]  Jagabathula,  S.,  et  al.:  Reputation-­based  worker  
filtering  in  crowdsourcing.  In:  Advances  in  Neural  Information  
Processing  Systems.  pp.  2492–2500  (2014)  
[KOS]  Karger,  D.R.,  Oh,  S.,  Shah,  D.:  Iterative  learning  for  reliable  
crowdsourcing  systems.  In:  Advances  in  neural  information  
processing  systems.  pp.  1953–1961  (2011)
Weighted  MV  Heuristic  
• We  use  weighted  MV  with  weights  equal  
to  reliabilities  of  volunteers
Reliability=  2  Pcorrect  answer
-­ 1
• The  heuristic  is  combined  with  iterative  
removal  of  a  volunteer  with  the  highest  
penalty [Hard  penalty]. Then  recalculate  
penalties,  and  obtain  new  results  for  the  
weighted  MV.  
Accuracy  for  ‘crowdsourcing’  algorithms
Baseline:    AdaBoost (35  features)  
91.08
Accuracy  for  ‘crowdsourcing’  
algorithms  with  image  thresholding.  
Only  images  with  
at  least  10  votes  
are  left  in  the  
expert  dataset:
404  images,  
1777  volunteers.  
Possible  explanations
Spammers
Malicious	
  
Annotators
Good	
  
Annotators
Biased	
  
Annotators
Biased
Annotators
ROCs  for  all    
1813  volunteers
on  the  expert  dataset
Raykar,  V.C.:  Eliminating  Spammers  and  Ranking  Annotators  for  Crowdsourced
Labeling  Tasks.  JMLR  13,  491–518  (2012)  
Volunteers’  ROCs
Spammers
Malicious	
  
Annotators
Good	
  
Annotators
Biased	
  
Annotators
Biased
Annotators
Threshold  =  12  votes,  262  volunteers.  
Volunteers’  ROCs
Spammers
Malicious	
  
Annotators
Good	
  
Annotators
Biased	
  
Annotators
Biased
Annotators
Threshold  =  100  votes, 24  volunteers.  
Accuracy
Due  to  image  processing  step
• accuracy  of  MV  increased  by  2%,  
• workload  decreased  by  6  %.
Using  threshold  for  voters  and  the  heuristic,  
we  increase  accuracy  to  95.5%.
Conclusions  
• Numerical  experiments  show  that  MV  performs  on  a  par  
with  all  algorithms.  
• Possible  explanation:  high  accuracy  of  frequently  voting  
volunteers  coupled  with  the  absence  of  spammers.
• Image  thresholding by  number  of  votes  helps  to  improve  
the  results  of  all  algorithms  similarly.
• To  summarize,  good  annotators  eliminate  any  advantages  
of  state  of  the  art  algorithms  over  MV.  
• Image  preprocessing  is  the  only  way  to  improve  accuracy!
Future  plans
• Development  of  theory
• More  case  studies  in  this  field,  hopefully  
with  spammers  and  malicious  voters!
• Elaborate  approach  to  ‘Maybe’  votes
The  baseline
Ø We  apply  SVD  to  the  whole  dataset  to  reduce  
dimensionality.  
Ø We  find  an  appropriate  choice  for  the  number  of  
features:  5,  14,  35.  
Ø Transform  the  feature  space  of  the  testing  and  
training  subsets  accordingly.  
Ø On  the  basis  of  10-­fold  cross-­validation  of  the  training  
subset,  we  fit  parameters  for  the  AdaBoost and  
Random  Forest  algorithms.  For  Linear  Discriminant  
Analysis  (LDA),  we  use  default  parameters.  
Ø The  accuracy  of  the  algorithms  with  fitted  parameters  
was  estimated  using  the  testing  subset.
Before and after merging of duplicates
Votes per image Unique votes
Repetition of images
(average)
Before [1 … 269] [1 … 88] 4
After [1…6042] [1…942] 13
900 volunteers
saw the same
image 7 times
Accuracy  for  ‘crowdsourcing’  
algorithms  with  image  thresholding.  
Only  images  with  
at  least  4  votes  
are  left  in  the  
expert  dataset:
729  images,  
1812  volunteers.  
Volunteers’  ROCs
Spammers
Malicious	
  
Annotators
Good	
  
Annotators
Biased	
  
Annotators
Biased
Annotators
Threshold  =  44  votes,  52 volunteers.  

More Related Content

What's hot

B045050812
B045050812B045050812
B045050812
IJERA Editor
 
Fuzzy Logic based Contrast Enhancement
Fuzzy Logic based Contrast EnhancementFuzzy Logic based Contrast Enhancement
Fuzzy Logic based Contrast Enhancement
Samrudh Keshava Kumar
 
Image Co-segmentation via Saliency Co-fusion
Image Co-segmentation via Saliency Co-fusionImage Co-segmentation via Saliency Co-fusion
Image Co-segmentation via Saliency Co-fusion
Koteswar Rao Jerripothula
 
Color reduction using the combination of the kohonen self organized feature m...
Color reduction using the combination of the kohonen self organized feature m...Color reduction using the combination of the kohonen self organized feature m...
Color reduction using the combination of the kohonen self organized feature m...
Konstantinos Zagoris
 
HCMUS at MediaEval 2020: Ensembles of Temporal Deep Neural Networks for Table...
HCMUS at MediaEval 2020: Ensembles of Temporal Deep Neural Networks for Table...HCMUS at MediaEval 2020: Ensembles of Temporal Deep Neural Networks for Table...
HCMUS at MediaEval 2020: Ensembles of Temporal Deep Neural Networks for Table...
multimediaeval
 
histogram equalization of grayscale and color image
histogram equalization of grayscale and color imagehistogram equalization of grayscale and color image
histogram equalization of grayscale and color image
Saeed Ullah
 
Btv thesis defense_v1.02-final
Btv thesis defense_v1.02-finalBtv thesis defense_v1.02-final
Btv thesis defense_v1.02-final
Vinh Bui
 
Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Anal...
Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Anal...Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Anal...
Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Anal...
JaeJun Yoo
 
Iccv2009 recognition and learning object categories p2 c01 - recognizing a ...
Iccv2009 recognition and learning object categories   p2 c01 - recognizing a ...Iccv2009 recognition and learning object categories   p2 c01 - recognizing a ...
Iccv2009 recognition and learning object categories p2 c01 - recognizing a ...
zukun
 
Mr image compression based on selection of mother wavelet and lifting based w...
Mr image compression based on selection of mother wavelet and lifting based w...Mr image compression based on selection of mother wavelet and lifting based w...
Mr image compression based on selection of mother wavelet and lifting based w...
ijma
 
Fractional step discriminant pruning
Fractional step discriminant pruningFractional step discriminant pruning
Fractional step discriminant pruning
VasileiosMezaris
 
International Journal of Engineering and Science Invention (IJESI)
International Journal of Engineering and Science Invention (IJESI)International Journal of Engineering and Science Invention (IJESI)
International Journal of Engineering and Science Invention (IJESI)
inventionjournals
 
Super Resolution of Image
Super Resolution of ImageSuper Resolution of Image
Super Resolution of Image
Satheesh K
 
poster
posterposter
TUPLE VALUE BASED MULTIPLICATIVE DATA PERTURBATION APPROACH TO PRESERVE PRIVA...
TUPLE VALUE BASED MULTIPLICATIVE DATA PERTURBATION APPROACH TO PRESERVE PRIVA...TUPLE VALUE BASED MULTIPLICATIVE DATA PERTURBATION APPROACH TO PRESERVE PRIVA...
TUPLE VALUE BASED MULTIPLICATIVE DATA PERTURBATION APPROACH TO PRESERVE PRIVA...
IJDKP
 
Image super resolution based on
Image super resolution based onImage super resolution based on
Image super resolution based on
jpstudcorner
 
Detecting image splicing in the wild Web
Detecting image splicing in the wild WebDetecting image splicing in the wild Web
Detecting image splicing in the wild Web
Symeon Papadopoulos
 
538 207-219
538 207-219538 207-219
538 207-219
idescitation
 
FAST ALGORITHMS FOR UNSUPERVISED LEARNING IN LARGE DATA SETS
FAST ALGORITHMS FOR UNSUPERVISED LEARNING IN LARGE DATA SETSFAST ALGORITHMS FOR UNSUPERVISED LEARNING IN LARGE DATA SETS
FAST ALGORITHMS FOR UNSUPERVISED LEARNING IN LARGE DATA SETS
csandit
 
Bayesian Networks with R and Hadoop
Bayesian Networks with R and HadoopBayesian Networks with R and Hadoop
Bayesian Networks with R and Hadoop
Ofer Mendelevitch
 

What's hot (20)

B045050812
B045050812B045050812
B045050812
 
Fuzzy Logic based Contrast Enhancement
Fuzzy Logic based Contrast EnhancementFuzzy Logic based Contrast Enhancement
Fuzzy Logic based Contrast Enhancement
 
Image Co-segmentation via Saliency Co-fusion
Image Co-segmentation via Saliency Co-fusionImage Co-segmentation via Saliency Co-fusion
Image Co-segmentation via Saliency Co-fusion
 
Color reduction using the combination of the kohonen self organized feature m...
Color reduction using the combination of the kohonen self organized feature m...Color reduction using the combination of the kohonen self organized feature m...
Color reduction using the combination of the kohonen self organized feature m...
 
HCMUS at MediaEval 2020: Ensembles of Temporal Deep Neural Networks for Table...
HCMUS at MediaEval 2020: Ensembles of Temporal Deep Neural Networks for Table...HCMUS at MediaEval 2020: Ensembles of Temporal Deep Neural Networks for Table...
HCMUS at MediaEval 2020: Ensembles of Temporal Deep Neural Networks for Table...
 
histogram equalization of grayscale and color image
histogram equalization of grayscale and color imagehistogram equalization of grayscale and color image
histogram equalization of grayscale and color image
 
Btv thesis defense_v1.02-final
Btv thesis defense_v1.02-finalBtv thesis defense_v1.02-final
Btv thesis defense_v1.02-final
 
Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Anal...
Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Anal...Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Anal...
Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Anal...
 
Iccv2009 recognition and learning object categories p2 c01 - recognizing a ...
Iccv2009 recognition and learning object categories   p2 c01 - recognizing a ...Iccv2009 recognition and learning object categories   p2 c01 - recognizing a ...
Iccv2009 recognition and learning object categories p2 c01 - recognizing a ...
 
Mr image compression based on selection of mother wavelet and lifting based w...
Mr image compression based on selection of mother wavelet and lifting based w...Mr image compression based on selection of mother wavelet and lifting based w...
Mr image compression based on selection of mother wavelet and lifting based w...
 
Fractional step discriminant pruning
Fractional step discriminant pruningFractional step discriminant pruning
Fractional step discriminant pruning
 
International Journal of Engineering and Science Invention (IJESI)
International Journal of Engineering and Science Invention (IJESI)International Journal of Engineering and Science Invention (IJESI)
International Journal of Engineering and Science Invention (IJESI)
 
Super Resolution of Image
Super Resolution of ImageSuper Resolution of Image
Super Resolution of Image
 
poster
posterposter
poster
 
TUPLE VALUE BASED MULTIPLICATIVE DATA PERTURBATION APPROACH TO PRESERVE PRIVA...
TUPLE VALUE BASED MULTIPLICATIVE DATA PERTURBATION APPROACH TO PRESERVE PRIVA...TUPLE VALUE BASED MULTIPLICATIVE DATA PERTURBATION APPROACH TO PRESERVE PRIVA...
TUPLE VALUE BASED MULTIPLICATIVE DATA PERTURBATION APPROACH TO PRESERVE PRIVA...
 
Image super resolution based on
Image super resolution based onImage super resolution based on
Image super resolution based on
 
Detecting image splicing in the wild Web
Detecting image splicing in the wild WebDetecting image splicing in the wild Web
Detecting image splicing in the wild Web
 
538 207-219
538 207-219538 207-219
538 207-219
 
FAST ALGORITHMS FOR UNSUPERVISED LEARNING IN LARGE DATA SETS
FAST ALGORITHMS FOR UNSUPERVISED LEARNING IN LARGE DATA SETSFAST ALGORITHMS FOR UNSUPERVISED LEARNING IN LARGE DATA SETS
FAST ALGORITHMS FOR UNSUPERVISED LEARNING IN LARGE DATA SETS
 
Bayesian Networks with R and Hadoop
Bayesian Networks with R and HadoopBayesian Networks with R and Hadoop
Bayesian Networks with R and Hadoop
 

Similar to Artem Baklanov - Votes Aggregation Techniques in Geo-Wiki Crowdsourcing Game: a Case Study

A multilevel automatic thresholding method based on a genetic algorithm for a...
A multilevel automatic thresholding method based on a genetic algorithm for a...A multilevel automatic thresholding method based on a genetic algorithm for a...
A multilevel automatic thresholding method based on a genetic algorithm for a...
Akshit Arora
 
Development and Comparison of Image Fusion Techniques for CT&MRI Images
Development and Comparison of Image Fusion Techniques for CT&MRI ImagesDevelopment and Comparison of Image Fusion Techniques for CT&MRI Images
Development and Comparison of Image Fusion Techniques for CT&MRI Images
IJERA Editor
 
EFFICIENT APPROACH FOR CONTENT BASED IMAGE RETRIEVAL USING MULTIPLE SVM IN YA...
EFFICIENT APPROACH FOR CONTENT BASED IMAGE RETRIEVAL USING MULTIPLE SVM IN YA...EFFICIENT APPROACH FOR CONTENT BASED IMAGE RETRIEVAL USING MULTIPLE SVM IN YA...
EFFICIENT APPROACH FOR CONTENT BASED IMAGE RETRIEVAL USING MULTIPLE SVM IN YA...
cscpconf
 
Efficient Approach for Content Based Image Retrieval Using Multiple SVM in YA...
Efficient Approach for Content Based Image Retrieval Using Multiple SVM in YA...Efficient Approach for Content Based Image Retrieval Using Multiple SVM in YA...
Efficient Approach for Content Based Image Retrieval Using Multiple SVM in YA...
csandit
 
Fuzzy entropy based optimal
Fuzzy entropy based optimalFuzzy entropy based optimal
Fuzzy entropy based optimal
ijsc
 
Generative Adversarial Networks and Their Applications in Medical Imaging
Generative Adversarial Networks  and Their Applications in Medical ImagingGenerative Adversarial Networks  and Their Applications in Medical Imaging
Generative Adversarial Networks and Their Applications in Medical Imaging
Sanghoon Hong
 
MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...
MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...
MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...
ijma
 
IMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHES
IMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHESIMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHES
IMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHES
Vikash Kumar
 
SINGLE IMAGE SUPER RESOLUTION: A COMPARATIVE STUDY
SINGLE IMAGE SUPER RESOLUTION: A COMPARATIVE STUDYSINGLE IMAGE SUPER RESOLUTION: A COMPARATIVE STUDY
SINGLE IMAGE SUPER RESOLUTION: A COMPARATIVE STUDY
csandit
 
AGE AND GENDER DETECTION.pptx
AGE AND GENDER DETECTION.pptxAGE AND GENDER DETECTION.pptx
AGE AND GENDER DETECTION.pptx
ssuserb4a9ba
 
ageandgenderdetection-220802061020-9ee5a2cd.pptx
ageandgenderdetection-220802061020-9ee5a2cd.pptxageandgenderdetection-220802061020-9ee5a2cd.pptx
ageandgenderdetection-220802061020-9ee5a2cd.pptx
dhaliwalharsh055
 
Sparse Sampling in Digital Image Processing
Sparse Sampling in Digital Image ProcessingSparse Sampling in Digital Image Processing
Sparse Sampling in Digital Image Processing
Eswar Publications
 
Using HOG Descriptors on Superpixels for Human Detection of UAV Imagery
Using HOG Descriptors on Superpixels for Human Detection of UAV ImageryUsing HOG Descriptors on Superpixels for Human Detection of UAV Imagery
Using HOG Descriptors on Superpixels for Human Detection of UAV Imagery
Wai Nwe Tun
 
An Evolutionary Dynamic Clustering based Colour Image Segmentation
An Evolutionary Dynamic Clustering based Colour Image SegmentationAn Evolutionary Dynamic Clustering based Colour Image Segmentation
An Evolutionary Dynamic Clustering based Colour Image Segmentation
CSCJournals
 
Learning from Computer Simulation to Tackle Real-World Problems
Learning from Computer Simulation to Tackle Real-World ProblemsLearning from Computer Simulation to Tackle Real-World Problems
Learning from Computer Simulation to Tackle Real-World Problems
NAVER Engineering
 
IRJET - Explicit Content Detection using Faster R-CNN and SSD Mobilenet V2
 IRJET - Explicit Content Detection using Faster R-CNN and SSD Mobilenet V2 IRJET - Explicit Content Detection using Faster R-CNN and SSD Mobilenet V2
IRJET - Explicit Content Detection using Faster R-CNN and SSD Mobilenet V2
IRJET Journal
 
Research Paper v2.0
Research Paper v2.0Research Paper v2.0
Research Paper v2.0
Kapil Tiwari
 
Ijetr011958
Ijetr011958Ijetr011958
Ijetr011958
ER Publication.org
 
ppt icitisee 2022_without_recording.pptx
ppt icitisee 2022_without_recording.pptxppt icitisee 2022_without_recording.pptx
ppt icitisee 2022_without_recording.pptx
ssusera4da91
 
自然方策勾配法の基礎と応用
自然方策勾配法の基礎と応用自然方策勾配法の基礎と応用
自然方策勾配法の基礎と応用
Ryo Iwaki
 

Similar to Artem Baklanov - Votes Aggregation Techniques in Geo-Wiki Crowdsourcing Game: a Case Study (20)

A multilevel automatic thresholding method based on a genetic algorithm for a...
A multilevel automatic thresholding method based on a genetic algorithm for a...A multilevel automatic thresholding method based on a genetic algorithm for a...
A multilevel automatic thresholding method based on a genetic algorithm for a...
 
Development and Comparison of Image Fusion Techniques for CT&MRI Images
Development and Comparison of Image Fusion Techniques for CT&MRI ImagesDevelopment and Comparison of Image Fusion Techniques for CT&MRI Images
Development and Comparison of Image Fusion Techniques for CT&MRI Images
 
EFFICIENT APPROACH FOR CONTENT BASED IMAGE RETRIEVAL USING MULTIPLE SVM IN YA...
EFFICIENT APPROACH FOR CONTENT BASED IMAGE RETRIEVAL USING MULTIPLE SVM IN YA...EFFICIENT APPROACH FOR CONTENT BASED IMAGE RETRIEVAL USING MULTIPLE SVM IN YA...
EFFICIENT APPROACH FOR CONTENT BASED IMAGE RETRIEVAL USING MULTIPLE SVM IN YA...
 
Efficient Approach for Content Based Image Retrieval Using Multiple SVM in YA...
Efficient Approach for Content Based Image Retrieval Using Multiple SVM in YA...Efficient Approach for Content Based Image Retrieval Using Multiple SVM in YA...
Efficient Approach for Content Based Image Retrieval Using Multiple SVM in YA...
 
Fuzzy entropy based optimal
Fuzzy entropy based optimalFuzzy entropy based optimal
Fuzzy entropy based optimal
 
Generative Adversarial Networks and Their Applications in Medical Imaging
Generative Adversarial Networks  and Their Applications in Medical ImagingGenerative Adversarial Networks  and Their Applications in Medical Imaging
Generative Adversarial Networks and Their Applications in Medical Imaging
 
MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...
MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...
MR Image Compression Based on Selection of Mother Wavelet and Lifting Based W...
 
IMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHES
IMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHESIMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHES
IMAGE CLASSIFICATION USING DIFFERENT CLASSICAL APPROACHES
 
SINGLE IMAGE SUPER RESOLUTION: A COMPARATIVE STUDY
SINGLE IMAGE SUPER RESOLUTION: A COMPARATIVE STUDYSINGLE IMAGE SUPER RESOLUTION: A COMPARATIVE STUDY
SINGLE IMAGE SUPER RESOLUTION: A COMPARATIVE STUDY
 
AGE AND GENDER DETECTION.pptx
AGE AND GENDER DETECTION.pptxAGE AND GENDER DETECTION.pptx
AGE AND GENDER DETECTION.pptx
 
ageandgenderdetection-220802061020-9ee5a2cd.pptx
ageandgenderdetection-220802061020-9ee5a2cd.pptxageandgenderdetection-220802061020-9ee5a2cd.pptx
ageandgenderdetection-220802061020-9ee5a2cd.pptx
 
Sparse Sampling in Digital Image Processing
Sparse Sampling in Digital Image ProcessingSparse Sampling in Digital Image Processing
Sparse Sampling in Digital Image Processing
 
Using HOG Descriptors on Superpixels for Human Detection of UAV Imagery
Using HOG Descriptors on Superpixels for Human Detection of UAV ImageryUsing HOG Descriptors on Superpixels for Human Detection of UAV Imagery
Using HOG Descriptors on Superpixels for Human Detection of UAV Imagery
 
An Evolutionary Dynamic Clustering based Colour Image Segmentation
An Evolutionary Dynamic Clustering based Colour Image SegmentationAn Evolutionary Dynamic Clustering based Colour Image Segmentation
An Evolutionary Dynamic Clustering based Colour Image Segmentation
 
Learning from Computer Simulation to Tackle Real-World Problems
Learning from Computer Simulation to Tackle Real-World ProblemsLearning from Computer Simulation to Tackle Real-World Problems
Learning from Computer Simulation to Tackle Real-World Problems
 
IRJET - Explicit Content Detection using Faster R-CNN and SSD Mobilenet V2
 IRJET - Explicit Content Detection using Faster R-CNN and SSD Mobilenet V2 IRJET - Explicit Content Detection using Faster R-CNN and SSD Mobilenet V2
IRJET - Explicit Content Detection using Faster R-CNN and SSD Mobilenet V2
 
Research Paper v2.0
Research Paper v2.0Research Paper v2.0
Research Paper v2.0
 
Ijetr011958
Ijetr011958Ijetr011958
Ijetr011958
 
ppt icitisee 2022_without_recording.pptx
ppt icitisee 2022_without_recording.pptxppt icitisee 2022_without_recording.pptx
ppt icitisee 2022_without_recording.pptx
 
自然方策勾配法の基礎と応用
自然方策勾配法の基礎と応用自然方策勾配法の基礎と応用
自然方策勾配法の基礎と応用
 

More from AIST

Alexey Mikhaylichenko - Automatic Detection of Bone Contours in X-Ray Images
Alexey Mikhaylichenko - Automatic Detection of Bone Contours in X-Ray  ImagesAlexey Mikhaylichenko - Automatic Detection of Bone Contours in X-Ray  Images
Alexey Mikhaylichenko - Automatic Detection of Bone Contours in X-Ray Images
AIST
 
Алена Ильина и Иван Бибилов, GoTo - GoTo школы, конкурсы и хакатоны
Алена Ильина и Иван Бибилов, GoTo - GoTo школы, конкурсы и хакатоныАлена Ильина и Иван Бибилов, GoTo - GoTo школы, конкурсы и хакатоны
Алена Ильина и Иван Бибилов, GoTo - GoTo школы, конкурсы и хакатоны
AIST
 
Станислав Кралин, Сайтсофт - Связанные открытые данные федеральных органов ис...
Станислав Кралин, Сайтсофт - Связанные открытые данные федеральных органов ис...Станислав Кралин, Сайтсофт - Связанные открытые данные федеральных органов ис...
Станислав Кралин, Сайтсофт - Связанные открытые данные федеральных органов ис...
AIST
 
Павел Браславский,Velpas - Velpas: мобильный визуальный поиск
Павел Браславский,Velpas - Velpas: мобильный визуальный поискПавел Браславский,Velpas - Velpas: мобильный визуальный поиск
Павел Браславский,Velpas - Velpas: мобильный визуальный поиск
AIST
 
Евгений Цымбалов, Webgames - Методы машинного обучения для задач игровой анал...
Евгений Цымбалов, Webgames - Методы машинного обучения для задач игровой анал...Евгений Цымбалов, Webgames - Методы машинного обучения для задач игровой анал...
Евгений Цымбалов, Webgames - Методы машинного обучения для задач игровой анал...
AIST
 
Александр Москвичев, EveResearch - Алгоритмы анализа данных в маркетинговых и...
Александр Москвичев, EveResearch - Алгоритмы анализа данных в маркетинговых и...Александр Москвичев, EveResearch - Алгоритмы анализа данных в маркетинговых и...
Александр Москвичев, EveResearch - Алгоритмы анализа данных в маркетинговых и...
AIST
 
Петр Ермаков, HeadHunter - Модерация резюме: от людей к роботам. Машинное обу...
Петр Ермаков, HeadHunter - Модерация резюме: от людей к роботам. Машинное обу...Петр Ермаков, HeadHunter - Модерация резюме: от людей к роботам. Машинное обу...
Петр Ермаков, HeadHunter - Модерация резюме: от людей к роботам. Машинное обу...
AIST
 
Иосиф Иткин, Exactpro - TBA
Иосиф Иткин, Exactpro - TBAИосиф Иткин, Exactpro - TBA
Иосиф Иткин, Exactpro - TBA
AIST
 
Nikolay Karpov - Evolvable Semantic Platform for Facilitating Knowledge Exchange
Nikolay Karpov - Evolvable Semantic Platform for Facilitating Knowledge ExchangeNikolay Karpov - Evolvable Semantic Platform for Facilitating Knowledge Exchange
Nikolay Karpov - Evolvable Semantic Platform for Facilitating Knowledge Exchange
AIST
 
George Moiseev - Classification of E-commerce Websites by Product Categories
George Moiseev - Classification of E-commerce Websites by Product CategoriesGeorge Moiseev - Classification of E-commerce Websites by Product Categories
George Moiseev - Classification of E-commerce Websites by Product Categories
AIST
 
Elena Bruches - The Hybrid Approach to Part-of-Speech Disambiguation
Elena Bruches - The Hybrid Approach to Part-of-Speech DisambiguationElena Bruches - The Hybrid Approach to Part-of-Speech Disambiguation
Elena Bruches - The Hybrid Approach to Part-of-Speech Disambiguation
AIST
 
Marina Danshina - The methodology of automated decryption of znamenny chants
Marina Danshina - The methodology of automated decryption of znamenny chantsMarina Danshina - The methodology of automated decryption of znamenny chants
Marina Danshina - The methodology of automated decryption of znamenny chants
AIST
 
Edward Klyshinsky - The Corpus of Syntactic Co-occurences: the First Glance
Edward Klyshinsky - The Corpus of Syntactic Co-occurences: the First GlanceEdward Klyshinsky - The Corpus of Syntactic Co-occurences: the First Glance
Edward Klyshinsky - The Corpus of Syntactic Co-occurences: the First Glance
AIST
 
Galina Lavrentyeva - Anti-spoofing Methods for Automatic Speaker Verification...
Galina Lavrentyeva - Anti-spoofing Methods for Automatic Speaker Verification...Galina Lavrentyeva - Anti-spoofing Methods for Automatic Speaker Verification...
Galina Lavrentyeva - Anti-spoofing Methods for Automatic Speaker Verification...
AIST
 
Oleksandr Frei and Murat Apishev - Parallel Non-blocking Deterministic Algori...
Oleksandr Frei and Murat Apishev - Parallel Non-blocking Deterministic Algori...Oleksandr Frei and Murat Apishev - Parallel Non-blocking Deterministic Algori...
Oleksandr Frei and Murat Apishev - Parallel Non-blocking Deterministic Algori...
AIST
 
Kaytoue Mehdi - Finding duplicate labels in behavioral data: an application f...
Kaytoue Mehdi - Finding duplicate labels in behavioral data: an application f...Kaytoue Mehdi - Finding duplicate labels in behavioral data: an application f...
Kaytoue Mehdi - Finding duplicate labels in behavioral data: an application f...
AIST
 
Valeri Labunets - The bichromatic excitable Schrodinger metamedium
Valeri Labunets - The bichromatic excitable Schrodinger metamediumValeri Labunets - The bichromatic excitable Schrodinger metamedium
Valeri Labunets - The bichromatic excitable Schrodinger metamedium
AIST
 
Valeri Labunets - Fast multiparametric wavelet transforms and packets for ima...
Valeri Labunets - Fast multiparametric wavelet transforms and packets for ima...Valeri Labunets - Fast multiparametric wavelet transforms and packets for ima...
Valeri Labunets - Fast multiparametric wavelet transforms and packets for ima...
AIST
 
Alexander Karkishchenko - Threefold Symmetry Detection in Hexagonal Images Ba...
Alexander Karkishchenko - Threefold Symmetry Detection in Hexagonal Images Ba...Alexander Karkishchenko - Threefold Symmetry Detection in Hexagonal Images Ba...
Alexander Karkishchenko - Threefold Symmetry Detection in Hexagonal Images Ba...
AIST
 
Artyom Makovetskii - An Efficient Algorithm for Total Variation Denoising
Artyom Makovetskii - An Efficient Algorithm for Total Variation DenoisingArtyom Makovetskii - An Efficient Algorithm for Total Variation Denoising
Artyom Makovetskii - An Efficient Algorithm for Total Variation Denoising
AIST
 

More from AIST (20)

Alexey Mikhaylichenko - Automatic Detection of Bone Contours in X-Ray Images
Alexey Mikhaylichenko - Automatic Detection of Bone Contours in X-Ray  ImagesAlexey Mikhaylichenko - Automatic Detection of Bone Contours in X-Ray  Images
Alexey Mikhaylichenko - Automatic Detection of Bone Contours in X-Ray Images
 
Алена Ильина и Иван Бибилов, GoTo - GoTo школы, конкурсы и хакатоны
Алена Ильина и Иван Бибилов, GoTo - GoTo школы, конкурсы и хакатоныАлена Ильина и Иван Бибилов, GoTo - GoTo школы, конкурсы и хакатоны
Алена Ильина и Иван Бибилов, GoTo - GoTo школы, конкурсы и хакатоны
 
Станислав Кралин, Сайтсофт - Связанные открытые данные федеральных органов ис...
Станислав Кралин, Сайтсофт - Связанные открытые данные федеральных органов ис...Станислав Кралин, Сайтсофт - Связанные открытые данные федеральных органов ис...
Станислав Кралин, Сайтсофт - Связанные открытые данные федеральных органов ис...
 
Павел Браславский,Velpas - Velpas: мобильный визуальный поиск
Павел Браславский,Velpas - Velpas: мобильный визуальный поискПавел Браславский,Velpas - Velpas: мобильный визуальный поиск
Павел Браславский,Velpas - Velpas: мобильный визуальный поиск
 
Евгений Цымбалов, Webgames - Методы машинного обучения для задач игровой анал...
Евгений Цымбалов, Webgames - Методы машинного обучения для задач игровой анал...Евгений Цымбалов, Webgames - Методы машинного обучения для задач игровой анал...
Евгений Цымбалов, Webgames - Методы машинного обучения для задач игровой анал...
 
Александр Москвичев, EveResearch - Алгоритмы анализа данных в маркетинговых и...
Александр Москвичев, EveResearch - Алгоритмы анализа данных в маркетинговых и...Александр Москвичев, EveResearch - Алгоритмы анализа данных в маркетинговых и...
Александр Москвичев, EveResearch - Алгоритмы анализа данных в маркетинговых и...
 
Петр Ермаков, HeadHunter - Модерация резюме: от людей к роботам. Машинное обу...
Петр Ермаков, HeadHunter - Модерация резюме: от людей к роботам. Машинное обу...Петр Ермаков, HeadHunter - Модерация резюме: от людей к роботам. Машинное обу...
Петр Ермаков, HeadHunter - Модерация резюме: от людей к роботам. Машинное обу...
 
Иосиф Иткин, Exactpro - TBA
Иосиф Иткин, Exactpro - TBAИосиф Иткин, Exactpro - TBA
Иосиф Иткин, Exactpro - TBA
 
Nikolay Karpov - Evolvable Semantic Platform for Facilitating Knowledge Exchange
Nikolay Karpov - Evolvable Semantic Platform for Facilitating Knowledge ExchangeNikolay Karpov - Evolvable Semantic Platform for Facilitating Knowledge Exchange
Nikolay Karpov - Evolvable Semantic Platform for Facilitating Knowledge Exchange
 
George Moiseev - Classification of E-commerce Websites by Product Categories
George Moiseev - Classification of E-commerce Websites by Product CategoriesGeorge Moiseev - Classification of E-commerce Websites by Product Categories
George Moiseev - Classification of E-commerce Websites by Product Categories
 
Elena Bruches - The Hybrid Approach to Part-of-Speech Disambiguation
Elena Bruches - The Hybrid Approach to Part-of-Speech DisambiguationElena Bruches - The Hybrid Approach to Part-of-Speech Disambiguation
Elena Bruches - The Hybrid Approach to Part-of-Speech Disambiguation
 
Marina Danshina - The methodology of automated decryption of znamenny chants
Marina Danshina - The methodology of automated decryption of znamenny chantsMarina Danshina - The methodology of automated decryption of znamenny chants
Marina Danshina - The methodology of automated decryption of znamenny chants
 
Edward Klyshinsky - The Corpus of Syntactic Co-occurences: the First Glance
Edward Klyshinsky - The Corpus of Syntactic Co-occurences: the First GlanceEdward Klyshinsky - The Corpus of Syntactic Co-occurences: the First Glance
Edward Klyshinsky - The Corpus of Syntactic Co-occurences: the First Glance
 
Galina Lavrentyeva - Anti-spoofing Methods for Automatic Speaker Verification...
Galina Lavrentyeva - Anti-spoofing Methods for Automatic Speaker Verification...Galina Lavrentyeva - Anti-spoofing Methods for Automatic Speaker Verification...
Galina Lavrentyeva - Anti-spoofing Methods for Automatic Speaker Verification...
 
Oleksandr Frei and Murat Apishev - Parallel Non-blocking Deterministic Algori...
Oleksandr Frei and Murat Apishev - Parallel Non-blocking Deterministic Algori...Oleksandr Frei and Murat Apishev - Parallel Non-blocking Deterministic Algori...
Oleksandr Frei and Murat Apishev - Parallel Non-blocking Deterministic Algori...
 
Kaytoue Mehdi - Finding duplicate labels in behavioral data: an application f...
Kaytoue Mehdi - Finding duplicate labels in behavioral data: an application f...Kaytoue Mehdi - Finding duplicate labels in behavioral data: an application f...
Kaytoue Mehdi - Finding duplicate labels in behavioral data: an application f...
 
Valeri Labunets - The bichromatic excitable Schrodinger metamedium
Valeri Labunets - The bichromatic excitable Schrodinger metamediumValeri Labunets - The bichromatic excitable Schrodinger metamedium
Valeri Labunets - The bichromatic excitable Schrodinger metamedium
 
Valeri Labunets - Fast multiparametric wavelet transforms and packets for ima...
Valeri Labunets - Fast multiparametric wavelet transforms and packets for ima...Valeri Labunets - Fast multiparametric wavelet transforms and packets for ima...
Valeri Labunets - Fast multiparametric wavelet transforms and packets for ima...
 
Alexander Karkishchenko - Threefold Symmetry Detection in Hexagonal Images Ba...
Alexander Karkishchenko - Threefold Symmetry Detection in Hexagonal Images Ba...Alexander Karkishchenko - Threefold Symmetry Detection in Hexagonal Images Ba...
Alexander Karkishchenko - Threefold Symmetry Detection in Hexagonal Images Ba...
 
Artyom Makovetskii - An Efficient Algorithm for Total Variation Denoising
Artyom Makovetskii - An Efficient Algorithm for Total Variation DenoisingArtyom Makovetskii - An Efficient Algorithm for Total Variation Denoising
Artyom Makovetskii - An Efficient Algorithm for Total Variation Denoising
 

Recently uploaded

Open Source Contributions to Postgres: The Basics POSETTE 2024
Open Source Contributions to Postgres: The Basics POSETTE 2024Open Source Contributions to Postgres: The Basics POSETTE 2024
Open Source Contributions to Postgres: The Basics POSETTE 2024
ElizabethGarrettChri
 
A presentation that explain the Power BI Licensing
A presentation that explain the Power BI LicensingA presentation that explain the Power BI Licensing
A presentation that explain the Power BI Licensing
AlessioFois2
 
06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM
06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM
06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM
Timothy Spann
 
The Ipsos - AI - Monitor 2024 Report.pdf
The  Ipsos - AI - Monitor 2024 Report.pdfThe  Ipsos - AI - Monitor 2024 Report.pdf
The Ipsos - AI - Monitor 2024 Report.pdf
Social Samosa
 
Everything you wanted to know about LIHTC
Everything you wanted to know about LIHTCEverything you wanted to know about LIHTC
Everything you wanted to know about LIHTC
Roger Valdez
 
ViewShift: Hassle-free Dynamic Policy Enforcement for Every Data Lake
ViewShift: Hassle-free Dynamic Policy Enforcement for Every Data LakeViewShift: Hassle-free Dynamic Policy Enforcement for Every Data Lake
ViewShift: Hassle-free Dynamic Policy Enforcement for Every Data Lake
Walaa Eldin Moustafa
 
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
apvysm8
 
原版一比一利兹贝克特大学毕业证(LeedsBeckett毕业证书)如何办理
原版一比一利兹贝克特大学毕业证(LeedsBeckett毕业证书)如何办理原版一比一利兹贝克特大学毕业证(LeedsBeckett毕业证书)如何办理
原版一比一利兹贝克特大学毕业证(LeedsBeckett毕业证书)如何办理
wyddcwye1
 
University of New South Wales degree offer diploma Transcript
University of New South Wales degree offer diploma TranscriptUniversity of New South Wales degree offer diploma Transcript
University of New South Wales degree offer diploma Transcript
soxrziqu
 
Beyond the Basics of A/B Tests: Highly Innovative Experimentation Tactics You...
Beyond the Basics of A/B Tests: Highly Innovative Experimentation Tactics You...Beyond the Basics of A/B Tests: Highly Innovative Experimentation Tactics You...
Beyond the Basics of A/B Tests: Highly Innovative Experimentation Tactics You...
Aggregage
 
Challenges of Nation Building-1.pptx with more important
Challenges of Nation Building-1.pptx with more importantChallenges of Nation Building-1.pptx with more important
Challenges of Nation Building-1.pptx with more important
Sm321
 
Predictably Improve Your B2B Tech Company's Performance by Leveraging Data
Predictably Improve Your B2B Tech Company's Performance by Leveraging DataPredictably Improve Your B2B Tech Company's Performance by Leveraging Data
Predictably Improve Your B2B Tech Company's Performance by Leveraging Data
Kiwi Creative
 
一比一原版(UO毕业证)渥太华大学毕业证如何办理
一比一原版(UO毕业证)渥太华大学毕业证如何办理一比一原版(UO毕业证)渥太华大学毕业证如何办理
一比一原版(UO毕业证)渥太华大学毕业证如何办理
aqzctr7x
 
Learn SQL from basic queries to Advance queries
Learn SQL from basic queries to Advance queriesLearn SQL from basic queries to Advance queries
Learn SQL from basic queries to Advance queries
manishkhaire30
 
一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理
一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理
一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理
nuttdpt
 
Intelligence supported media monitoring in veterinary medicine
Intelligence supported media monitoring in veterinary medicineIntelligence supported media monitoring in veterinary medicine
Intelligence supported media monitoring in veterinary medicine
AndrzejJarynowski
 
Experts live - Improving user adoption with AI
Experts live - Improving user adoption with AIExperts live - Improving user adoption with AI
Experts live - Improving user adoption with AI
jitskeb
 
一比一原版(Unimelb毕业证书)墨尔本大学毕业证如何办理
一比一原版(Unimelb毕业证书)墨尔本大学毕业证如何办理一比一原版(Unimelb毕业证书)墨尔本大学毕业证如何办理
一比一原版(Unimelb毕业证书)墨尔本大学毕业证如何办理
xclpvhuk
 
Palo Alto Cortex XDR presentation .......
Palo Alto Cortex XDR presentation .......Palo Alto Cortex XDR presentation .......
Palo Alto Cortex XDR presentation .......
Sachin Paul
 
一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理
一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理
一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理
nuttdpt
 

Recently uploaded (20)

Open Source Contributions to Postgres: The Basics POSETTE 2024
Open Source Contributions to Postgres: The Basics POSETTE 2024Open Source Contributions to Postgres: The Basics POSETTE 2024
Open Source Contributions to Postgres: The Basics POSETTE 2024
 
A presentation that explain the Power BI Licensing
A presentation that explain the Power BI LicensingA presentation that explain the Power BI Licensing
A presentation that explain the Power BI Licensing
 
06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM
06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM
06-12-2024-BudapestDataForum-BuildingReal-timePipelineswithFLaNK AIM
 
The Ipsos - AI - Monitor 2024 Report.pdf
The  Ipsos - AI - Monitor 2024 Report.pdfThe  Ipsos - AI - Monitor 2024 Report.pdf
The Ipsos - AI - Monitor 2024 Report.pdf
 
Everything you wanted to know about LIHTC
Everything you wanted to know about LIHTCEverything you wanted to know about LIHTC
Everything you wanted to know about LIHTC
 
ViewShift: Hassle-free Dynamic Policy Enforcement for Every Data Lake
ViewShift: Hassle-free Dynamic Policy Enforcement for Every Data LakeViewShift: Hassle-free Dynamic Policy Enforcement for Every Data Lake
ViewShift: Hassle-free Dynamic Policy Enforcement for Every Data Lake
 
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
办(uts毕业证书)悉尼科技大学毕业证学历证书原版一模一样
 
原版一比一利兹贝克特大学毕业证(LeedsBeckett毕业证书)如何办理
原版一比一利兹贝克特大学毕业证(LeedsBeckett毕业证书)如何办理原版一比一利兹贝克特大学毕业证(LeedsBeckett毕业证书)如何办理
原版一比一利兹贝克特大学毕业证(LeedsBeckett毕业证书)如何办理
 
University of New South Wales degree offer diploma Transcript
University of New South Wales degree offer diploma TranscriptUniversity of New South Wales degree offer diploma Transcript
University of New South Wales degree offer diploma Transcript
 
Beyond the Basics of A/B Tests: Highly Innovative Experimentation Tactics You...
Beyond the Basics of A/B Tests: Highly Innovative Experimentation Tactics You...Beyond the Basics of A/B Tests: Highly Innovative Experimentation Tactics You...
Beyond the Basics of A/B Tests: Highly Innovative Experimentation Tactics You...
 
Challenges of Nation Building-1.pptx with more important
Challenges of Nation Building-1.pptx with more importantChallenges of Nation Building-1.pptx with more important
Challenges of Nation Building-1.pptx with more important
 
Predictably Improve Your B2B Tech Company's Performance by Leveraging Data
Predictably Improve Your B2B Tech Company's Performance by Leveraging DataPredictably Improve Your B2B Tech Company's Performance by Leveraging Data
Predictably Improve Your B2B Tech Company's Performance by Leveraging Data
 
一比一原版(UO毕业证)渥太华大学毕业证如何办理
一比一原版(UO毕业证)渥太华大学毕业证如何办理一比一原版(UO毕业证)渥太华大学毕业证如何办理
一比一原版(UO毕业证)渥太华大学毕业证如何办理
 
Learn SQL from basic queries to Advance queries
Learn SQL from basic queries to Advance queriesLearn SQL from basic queries to Advance queries
Learn SQL from basic queries to Advance queries
 
一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理
一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理
一比一原版(UCSF文凭证书)旧金山分校毕业证如何办理
 
Intelligence supported media monitoring in veterinary medicine
Intelligence supported media monitoring in veterinary medicineIntelligence supported media monitoring in veterinary medicine
Intelligence supported media monitoring in veterinary medicine
 
Experts live - Improving user adoption with AI
Experts live - Improving user adoption with AIExperts live - Improving user adoption with AI
Experts live - Improving user adoption with AI
 
一比一原版(Unimelb毕业证书)墨尔本大学毕业证如何办理
一比一原版(Unimelb毕业证书)墨尔本大学毕业证如何办理一比一原版(Unimelb毕业证书)墨尔本大学毕业证如何办理
一比一原版(Unimelb毕业证书)墨尔本大学毕业证如何办理
 
Palo Alto Cortex XDR presentation .......
Palo Alto Cortex XDR presentation .......Palo Alto Cortex XDR presentation .......
Palo Alto Cortex XDR presentation .......
 
一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理
一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理
一比一原版(UCSB文凭证书)圣芭芭拉分校毕业证如何办理
 

Artem Baklanov - Votes Aggregation Techniques in Geo-Wiki Crowdsourcing Game: a Case Study

  • 1. Vote Aggregation Techniques in the Geo-Wiki Crowdsourcing Game: A Case Study Michael Khachay, Oleg Nurmukhametov Artem Baklanov Krasovskii Institute of Mathematics and Mechanics Postdoctoral Research Scholar, ASA, IIASA Krasovskii Institute of Mathematics and Mechanics Steffen Fritz, Carl Salk, Linda See, and Dmitry Shchepashchenko IIASA
  • 2. Young  Scientists  Summer  Program Since  1977,  IIASA’s  annual  3-­month  Young  Scientists   Summer  Program  (YSSP)  offers  opportunities  to   talented  young  researchers.
  • 3. Crowdsourcing = Crowd + Outsourcing Crowdsourcing is a new approach to perform tasks, when a group of distributed worldwide people in total can substitute an expert. Example: GEO-wiki project, ESM, IIASA
  • 4. Geo-Wiki: Cropland Capture Game 5 millions votes 170 000 satellite images 3000 volunteers 6 months Goal: land cover map www.geo-wiki.org d
  • 5. Preprocessing of images Blur detection Duplicate detection Benchmarking of algorithms for vote aggregation Outline of presentation Noise reduction Prediction
  • 6. Preprocessing of images Duplicate detection Blur detection Benchmarking of algorithms for vote aggregation Outline of presentation Noise reduction Prediction
  • 7. Blur detection algorithm [H Tang, 2012]. Input: image Output: coefficient of blur in [0, 100] Detection of blurry images 100%0%
  • 9. Blurry images 80% 90% 95% 98% 99%
  • 10. The  table  and  figure  are  from Tong,  Hanghang,  et  al.  "Blur  detection  for  digital   images  using  wavelet  transform." Multimedia   and  Expo,  2004.  ICME'04.  2004  IEEE   International  Conference  on.  Vol.  1.  IEEE,  2004. Blur detection algorithm
  • 11. A  key  idea • A= a  number  of  Roof-­Structure  and   Gstep-­Structure  edge  points  that  have  lost   their  sharpness. • B=  a  number  of  Roof-­Structure  and   Gstep-­Structure  edge  points.   • Blur  coefficient  =  100*A/B
  • 12. Detection of blurry images Results Current campaign: Noise reduction in dataset. + for future campaigns: Less workload, more joy. Blur coefficient Number of images 0% (Excellent images) 74 000 (38%) 80% - 100% (Extremely bad images) 3 200 (2%) Extremely bad images were removed
  • 13. Detection of duplicates Web Link Image unique identifier Link to image 1 http://cg.tuwien.ac.at/~sturn/crop/img_78.1875_52.9958_1000.jpg 2 http://cg.tuwien.ac.at/~sturn/crop/img_78.1875_52.9958_500.jpg 3 http://cg.tuwien.ac.at/~sturn/crop/img_78.1875_52.9958_300.jpg … … Binary file Image content
  • 14. Detection of duplicates Find the 10 differences
  • 15. Detection of duplicates The pixel-by-pixel differences Links: different Binary files: different Image content: similar
  • 16. pHash (perceptual hash) [Zauner, 2010]. Detection of duplicates
  • 17. pHash.  High  level  description 1. Resize  to  32x32  and  grayscale;; 2. Perform  DCT-­2  transformation  and  keep   top-­left  8x8;; 3. Calculate  the  average  value  of   alternating  components;; 4. Set  the  64  hash  bits  to  0  (or  1)  on   whether   DCT  values  more  (or  less)  the   average  value.
  • 18. Detection of copies Method Copies count Time to completion Binary data comparison (MD5) 2 700 Fast (~10 min) Image content comparison (pHash) 10 000 (6%) Slow(~7 hours) Results Current campaign: Increase of statistically significance; Reduction of dimensionality. + for future campaigns: Less workload, more joy. Votes  were  merged.
  • 19. Preprocessing of images Duplicate detection Blur detection Benchmarking of algorithms for vote aggregation Outline of presentation Noise reduction Prediction
  • 20. A  part  of  the  dataset was  annotated  by  an   expert  after  the  campaign  took  place: • 854 images;; • 1813 volunteers;; • 16  940 votes.   We    sample  two  subsets  for  training  and  testing   70/30 ratio Expert dataset
  • 21. Baseline  algorithms   We  use  SVD  to  reduce  dimensionality  and   10-­fold  cross-­validation  to  fit  parameters
  • 22. A  publicly  available  code  implements  KOS and  EM algorithm;;   both  are  implemented  in  conjunction  with  reputation   algorithm  (also  called  Hard  penalty).   https://github.com/ashwin90/Penalty-­based-­clustering   [EM]  Dawid,  A.P.,  Skene,  A.M.:  Maximum  likelihood  estimation  of   observer  error-­rates  using  the  em algorithm.  Applied  statistics  pp.   20–28  (1979) [Hard  penalty]  Jagabathula,  S.,  et  al.:  Reputation-­based  worker   filtering  in  crowdsourcing.  In:  Advances  in  Neural  Information   Processing  Systems.  pp.  2492–2500  (2014)   [KOS]  Karger,  D.R.,  Oh,  S.,  Shah,  D.:  Iterative  learning  for  reliable   crowdsourcing  systems.  In:  Advances  in  neural  information   processing  systems.  pp.  1953–1961  (2011)
  • 23. Weighted  MV  Heuristic   • We  use  weighted  MV  with  weights  equal   to  reliabilities  of  volunteers Reliability=  2  Pcorrect  answer -­ 1 • The  heuristic  is  combined  with  iterative   removal  of  a  volunteer  with  the  highest   penalty [Hard  penalty]. Then  recalculate   penalties,  and  obtain  new  results  for  the   weighted  MV.  
  • 24. Accuracy  for  ‘crowdsourcing’  algorithms Baseline:    AdaBoost (35  features)   91.08
  • 25. Accuracy  for  ‘crowdsourcing’   algorithms  with  image  thresholding.   Only  images  with   at  least  10  votes   are  left  in  the   expert  dataset: 404  images,   1777  volunteers.  
  • 26. Possible  explanations Spammers Malicious   Annotators Good   Annotators Biased   Annotators Biased Annotators ROCs  for  all     1813  volunteers on  the  expert  dataset Raykar,  V.C.:  Eliminating  Spammers  and  Ranking  Annotators  for  Crowdsourced Labeling  Tasks.  JMLR  13,  491–518  (2012)  
  • 27. Volunteers’  ROCs Spammers Malicious   Annotators Good   Annotators Biased   Annotators Biased Annotators Threshold  =  12  votes,  262  volunteers.  
  • 28. Volunteers’  ROCs Spammers Malicious   Annotators Good   Annotators Biased   Annotators Biased Annotators Threshold  =  100  votes, 24  volunteers.  
  • 29. Accuracy Due  to  image  processing  step • accuracy  of  MV  increased  by  2%,   • workload  decreased  by  6  %. Using  threshold  for  voters  and  the  heuristic,   we  increase  accuracy  to  95.5%.
  • 30. Conclusions   • Numerical  experiments  show  that  MV  performs  on  a  par   with  all  algorithms.   • Possible  explanation:  high  accuracy  of  frequently  voting   volunteers  coupled  with  the  absence  of  spammers. • Image  thresholding by  number  of  votes  helps  to  improve   the  results  of  all  algorithms  similarly. • To  summarize,  good  annotators  eliminate  any  advantages   of  state  of  the  art  algorithms  over  MV.   • Image  preprocessing  is  the  only  way  to  improve  accuracy!
  • 31. Future  plans • Development  of  theory • More  case  studies  in  this  field,  hopefully   with  spammers  and  malicious  voters! • Elaborate  approach  to  ‘Maybe’  votes
  • 32.
  • 33. The  baseline Ø We  apply  SVD  to  the  whole  dataset  to  reduce   dimensionality.   Ø We  find  an  appropriate  choice  for  the  number  of   features:  5,  14,  35.   Ø Transform  the  feature  space  of  the  testing  and   training  subsets  accordingly.   Ø On  the  basis  of  10-­fold  cross-­validation  of  the  training   subset,  we  fit  parameters  for  the  AdaBoost and   Random  Forest  algorithms.  For  Linear  Discriminant   Analysis  (LDA),  we  use  default  parameters.   Ø The  accuracy  of  the  algorithms  with  fitted  parameters   was  estimated  using  the  testing  subset.
  • 34. Before and after merging of duplicates Votes per image Unique votes Repetition of images (average) Before [1 … 269] [1 … 88] 4 After [1…6042] [1…942] 13 900 volunteers saw the same image 7 times
  • 35. Accuracy  for  ‘crowdsourcing’   algorithms  with  image  thresholding.   Only  images  with   at  least  4  votes   are  left  in  the   expert  dataset: 729  images,   1812  volunteers.  
  • 36. Volunteers’  ROCs Spammers Malicious   Annotators Good   Annotators Biased   Annotators Biased Annotators Threshold  =  44  votes,  52 volunteers.