SlideShare a Scribd company logo
1 of 39
Automatic Photo Selection for  Media and Entertainment  Applications Ekaterina Potapova,  Marta Egorova,  Ilia Safonov National Nuclear Research University  MEPhI  Moscow, Russia GraphiCon 2009 5-9 October
Applications Automatic Photo Selection for Media and Entertainment Applications GraphiCon 2009 2
Applications Automatic Photo Selection for Media and Entertainment Applications GraphiCon 2009 2
Applications – photo book Images are taken from printbook.ru, ehow.com, snapfish.com.au, smilebooks.co.uk GraphiCon 2009 3 Automatic Photo Selection for Media and Entertainment Applications
Applications – slide show Photos from ITaS’2008 GraphiCon 2009 4 Automatic Photo Selection for Media and Entertainment Applications
General workflow GraphiCon 2009 5 Automatic Photo Selection for Media and Entertainment Applications
GraphiCon 2009 5 Automatic Photo Selection for Media and Entertainment Applications General workflow Detection of low-quality photos
General workflow GraphiCon 2009 5 Automatic Photo Selection for Media and Entertainment Applications Detection of low-quality photos Adaptive quantization on time-camera plane
General workflow GraphiCon 2009 5 Automatic Photo Selection for Media and Entertainment Applications Selection of appealing photos Detection of low-quality photos Adaptive quantization on time-camera plane
Detection of low-quality photos GraphiCon 2009 6 Automatic Photo Selection for Media and Entertainment Applications
Estimation of JPEG quality A.Foi et al.,2007 Images are taken from en.wikipedia.org Quantization Table GraphiCon 2009 7 Automatic Photo Selection for Media and Entertainment Applications
Detection of backlit, low-contrast & blurred photos Two Ada Boost classifiers committee:  -for detection of low-contrast and backlit photos -for detection of blurred photos GraphiCon 2009 8 Automatic Photo Selection for Media and Entertainment Applications + Good photo Bad photo True False … …
Detection of backlit and low-contrast photos  - 1   S1/S2 -  ratio of tones in shadows to midtones GraphiCon 2009 9 Automatic Photo Selection for Media and Entertainment Applications
S11/S12  - ratio of tones in first to second part of shadows Detection of backlit and low-contrast photos  - 1   GraphiCon 2009 9 Automatic Photo Selection for Media and Entertainment Applications
M1/M2  -  ratio of the histogram maximum in shadows to the maximum in midtones Detection of backlit and low-contrast photos  - 1   GraphiCon 2009 9 Automatic Photo Selection for Media and Entertainment Applications
P1   - location of the histogram maximum in shadows P1 Detection of backlit and low-contrast photos  - 1   GraphiCon 2009 9 Automatic Photo Selection for Media and Entertainment Applications
C  –  global contrast H 0 C 0 C 1 H 1 Detection of backlit and low-contrast photos  - 1   GraphiCon 2009 9 Automatic Photo Selection for Media and Entertainment Applications
Training set:  480 photos Error rate on cross-validation test :  ~0.055 Testing set:  1830 with 2% affected by backlit and low-contrast photos The number of  False Positives  (FP) is 10  The number of  False Negatives  (FN) is 3  Low-contrast photo Backlit photo Detection of backlit and low-contrast photos  -  2  GraphiCon 2009 10 Automatic Photo Selection for Media and Entertainment Applications
Image Intensity image Z 1 =[-1 1] Z 2 =[-1  0 1] Z 3 =[-1  0 0 1] Z 10 =[-1  0 0 0 0 0 0 0 0 0 1] I.Safonov et al.,2008 … Edge image Histogram Normalized entropy Entropy to [0, 1] ? ? ? ? An An GraphiCon 2009 11 Detection of blurred photos Automatic Photo Selection for Media and Entertainment Applications
Crete et al., 2007 F.Crete et al.,2007 ? Image Blurred image Edge image Edge image Comparison of the  images HPF=[1 -1] LPF=[1 1 1 1 1 1 1 1 1]/9 Detection of blurred photos GraphiCon 2009 11 Automatic Photo Selection for Media and Entertainment Applications
Training set:  416 photos Error rate on cross-validation test :  ~0.07 Testing set:  1830 with 171 blurred photos The number of  False Positives  (FP) is 34 The number of  False Negatives  (FN) is 10 Detection of blurred photos GraphiCon 2009 11 Automatic Photo Selection for Media and Entertainment Applications
Time and camera-based quantization i  is an index of source   L  is time between the least and the most time for the largest source   Nps  is a number of sources   H = L/M   M  is count of images GraphiCon 2009 11 Automatic Photo Selection for Media and Entertainment Applications N region  < N group N region  < M Calculation of bounding boxes Partition into 2 app. equal subregions Seeking for the biggest region 1200 3600 2400 7200 0 36000 T, s 21600
GraphiCon 2009 12 Automatic Photo Selection for Media and Entertainment Applications Salient Photo Selection The most appealing photo is the most salient photo  L.Itti, C.Koch et al. Images are taken from the Internet
Conspicuity  maps Gaussian pyramids Image Intensity image r-channel g-channel b-channel R-channel G-channel B-channel Y-channel Orientation  map Intensity map Color map Saliency map Feature maps Gabor   pyramids GraphiCon 2009 13 Automatic Photo Selection for Media and Entertainment Applications Salient Photo Selection
original image saliency map intensity map color map orientation map ROI Automatic Photo Selection for Media and Entertainment Applications Salient Photo Selection GraphiCon 2009 14 Image is taken from the Internet
Automatic Photo Selection for Media and Entertainment Applications Salient Photo Selection GraphiCon 2009 15 124 88 11 100 81 92 62 83 105 70 Saliency Index
Automatic Photo Selection for Media and Entertainment Applications Salient Photo Selection GraphiCon 2009 15 83 11 124 Saliency Index 81 88 62 92 105 70 100
[object Object],[object Object],We consider, that images of people attracts more attention ,[object Object],Six places were detected erroneously ,[object Object],[object Object],[object Object],P.Viola, M.Jones, 2001 Automatic Photo Selection for Media and Entertainment Applications Face Detection GraphiCon 2009 16 Viola-Jones, Intel OpenCV Before modifications After modifications
Photos ranking Heuristic formula, experiments have shown that value w=25 gives the best result Automatic Photo Selection for Media and Entertainment Applications GraphiCon 2009 17 124 88 11 116 92 118 148 95 62 100
Photos ranking Heuristic formula, experiments have shown that value w=25 gives the best result Automatic Photo Selection for Media and Entertainment Applications GraphiCon 2009 17 118 62 124 88 11 100 116 92 148 95
Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 18
Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 18 Autocollage choice Our choice
Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 18
Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 18 Autocollage choice Our choice
Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 18
Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 18 Autocollage choice Our choice
Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 19 Proposed AutoCollage Random 14 1 4 3 3 3 Unacceptable 21 5 2 4 5 5 Acceptable 15 4 4 3 2 2 Agree with experts 9 1 4 1 1 2 Unacceptable 24 4 0 7 7 6 Acceptable 17 5 6 2 2 2 Agree with experts 4 1 1 0 1 1 Unacceptable 17 2 4 4 4 3 Acceptable 29 7 5 6 5 6 Agree with experts Sum Set 5 Set 4 Set 3 Set 2 Set 1
? Automatic Photo Selection for Media and Entertainment Applications Questions & Answers GraphiCon 2009 8
Automatic Photo Selection for Media and Entertainment Applications GraphiCon 2009 9 Thank you for your attention =)

More Related Content

Viewers also liked

San prohibited pesticide list april 2009
San prohibited pesticide list april 2009San prohibited pesticide list april 2009
San prohibited pesticide list april 2009Hung Pham Thai
 
Beatles -the_complete_songbook
Beatles  -the_complete_songbookBeatles  -the_complete_songbook
Beatles -the_complete_songbookHung Pham Thai
 
Financial market k42-2005
Financial market k42-2005Financial market k42-2005
Financial market k42-2005Hung Pham Thai
 
Ra saas certification manual aug 09
Ra saas certification manual aug 09Ra saas certification manual aug 09
Ra saas certification manual aug 09Hung Pham Thai
 
Technology Integration
Technology IntegrationTechnology Integration
Technology Integrationguest5b9bf4
 
Technology In The Classroom
Technology In The ClassroomTechnology In The Classroom
Technology In The Classroomhales4
 
Binh phong ma co dien toan tap 2 (china chess)
Binh phong ma co dien toan tap 2 (china chess)Binh phong ma co dien toan tap 2 (china chess)
Binh phong ma co dien toan tap 2 (china chess)Hung Pham Thai
 
Technology In The Classroom
Technology In The ClassroomTechnology In The Classroom
Technology In The Classroomhales4
 
Nicole's Technology Experience
Nicole's Technology ExperienceNicole's Technology Experience
Nicole's Technology Experiencehales4
 
16 Kynangquanlyxungdot
16 Kynangquanlyxungdot16 Kynangquanlyxungdot
16 KynangquanlyxungdotHung Pham Thai
 
Technology Integration
Technology IntegrationTechnology Integration
Technology Integrationcharlamt
 
Chinh Sach Cua Cong Ty
Chinh Sach Cua Cong TyChinh Sach Cua Cong Ty
Chinh Sach Cua Cong TyHung Pham Thai
 

Viewers also liked (20)

San prohibited pesticide list april 2009
San prohibited pesticide list april 2009San prohibited pesticide list april 2009
San prohibited pesticide list april 2009
 
Beatles -the_complete_songbook
Beatles  -the_complete_songbookBeatles  -the_complete_songbook
Beatles -the_complete_songbook
 
Financial market k42-2005
Financial market k42-2005Financial market k42-2005
Financial market k42-2005
 
Foundation k42-2005
Foundation k42-2005Foundation k42-2005
Foundation k42-2005
 
Ra saas certification manual aug 09
Ra saas certification manual aug 09Ra saas certification manual aug 09
Ra saas certification manual aug 09
 
Miss viet nam
Miss viet namMiss viet nam
Miss viet nam
 
Technology Integration
Technology IntegrationTechnology Integration
Technology Integration
 
Journal Summary
Journal SummaryJournal Summary
Journal Summary
 
Technology In The Classroom
Technology In The ClassroomTechnology In The Classroom
Technology In The Classroom
 
THU HOẠCH
THU HOẠCHTHU HOẠCH
THU HOẠCH
 
Binh phong ma co dien toan tap 2 (china chess)
Binh phong ma co dien toan tap 2 (china chess)Binh phong ma co dien toan tap 2 (china chess)
Binh phong ma co dien toan tap 2 (china chess)
 
Technology In The Classroom
Technology In The ClassroomTechnology In The Classroom
Technology In The Classroom
 
Today Is Tuesday
Today Is TuesdayToday Is Tuesday
Today Is Tuesday
 
CHỌN GIỐNG
CHỌN GIỐNGCHỌN GIỐNG
CHỌN GIỐNG
 
Nicole's Technology Experience
Nicole's Technology ExperienceNicole's Technology Experience
Nicole's Technology Experience
 
16 Kynangquanlyxungdot
16 Kynangquanlyxungdot16 Kynangquanlyxungdot
16 Kynangquanlyxungdot
 
Introduction to Linq
Introduction to LinqIntroduction to Linq
Introduction to Linq
 
Technology Integration
Technology IntegrationTechnology Integration
Technology Integration
 
Access vba 052009
Access vba 052009Access vba 052009
Access vba 052009
 
Chinh Sach Cua Cong Ty
Chinh Sach Cua Cong TyChinh Sach Cua Cong Ty
Chinh Sach Cua Cong Ty
 

Similar to Automatic Photo Selection For Media And Entertainment Applications

Image processing python
Image processing pythonImage processing python
Image processing pythonYAZIDI Imran
 
“Using an ISP for Real-time Data Augmentation,” a Presentation from Pony.AI
“Using an ISP for Real-time Data Augmentation,” a Presentation from Pony.AI“Using an ISP for Real-time Data Augmentation,” a Presentation from Pony.AI
“Using an ISP for Real-time Data Augmentation,” a Presentation from Pony.AIEdge AI and Vision Alliance
 
Tracking Robustness and Green View Index Estimation of Augmented and Diminish...
Tracking Robustness and Green View Index Estimation of Augmented and Diminish...Tracking Robustness and Green View Index Estimation of Augmented and Diminish...
Tracking Robustness and Green View Index Estimation of Augmented and Diminish...Tomohiro Fukuda
 
Smart Control of Traffic Signal System using Image Processing
Smart Control of Traffic Signal System using Image Processing Smart Control of Traffic Signal System using Image Processing
Smart Control of Traffic Signal System using Image Processing Raihan Bin-Mofidul
 
Applications of Image Processing and Real-Time embedded Systems in Autonomous...
Applications of Image Processing and Real-Time embedded Systems in Autonomous...Applications of Image Processing and Real-Time embedded Systems in Autonomous...
Applications of Image Processing and Real-Time embedded Systems in Autonomous...CSCJournals
 
TraVis CTTHES3
TraVis CTTHES3TraVis CTTHES3
TraVis CTTHES3Ni Aguirre
 
Quicklook technology assessment topmod software_ccchittim
Quicklook technology assessment topmod software_ccchittimQuicklook technology assessment topmod software_ccchittim
Quicklook technology assessment topmod software_ccchittimClaudia Chittim
 
Digital Image Forensics: camera fingerprint and its robustness
Digital Image Forensics: camera fingerprint and its robustness Digital Image Forensics: camera fingerprint and its robustness
Digital Image Forensics: camera fingerprint and its robustness Francesco Forestieri
 
Panorama Technique for 3D Animation movie, Design and Evaluating
Panorama Technique for 3D Animation movie, Design and EvaluatingPanorama Technique for 3D Animation movie, Design and Evaluating
Panorama Technique for 3D Animation movie, Design and EvaluatingIOSRjournaljce
 
IRJET- Reckoning the Vehicle using MATLAB
IRJET- Reckoning the Vehicle using MATLABIRJET- Reckoning the Vehicle using MATLAB
IRJET- Reckoning the Vehicle using MATLABIRJET Journal
 
Capturing Real-World Materials for Real-Time Development - Unite LA
Capturing Real-World Materials for Real-Time Development - Unite LACapturing Real-World Materials for Real-Time Development - Unite LA
Capturing Real-World Materials for Real-Time Development - Unite LAUnity Technologies
 
Face in video evaluation (five)
Face in video evaluation (five)Face in video evaluation (five)
Face in video evaluation (five)Sungkwan Park
 
Gutter Detection and Ranging
Gutter Detection and RangingGutter Detection and Ranging
Gutter Detection and RangingBEN ROSE
 
Rapid Laser Scanning the process
Rapid Laser Scanning the processRapid Laser Scanning the process
Rapid Laser Scanning the processSeeview Solutions
 
Sample 2018 top 5 3 d time-of-flight image sensor players in north america, e...
Sample 2018 top 5 3 d time-of-flight image sensor players in north america, e...Sample 2018 top 5 3 d time-of-flight image sensor players in north america, e...
Sample 2018 top 5 3 d time-of-flight image sensor players in north america, e...俊豪 馬
 
APPLICATION OF IP TECHNIQUES IN TRAFFIC CONTROL SYSTEM
APPLICATION OF IP TECHNIQUES IN TRAFFIC CONTROL SYSTEMAPPLICATION OF IP TECHNIQUES IN TRAFFIC CONTROL SYSTEM
APPLICATION OF IP TECHNIQUES IN TRAFFIC CONTROL SYSTEMAshik Ask
 
IRJET- Dynamic Traffic Management System
IRJET- Dynamic Traffic Management SystemIRJET- Dynamic Traffic Management System
IRJET- Dynamic Traffic Management SystemIRJET Journal
 
Photosynth
PhotosynthPhotosynth
Photosynthsemsem23
 

Similar to Automatic Photo Selection For Media And Entertainment Applications (20)

Image processing python
Image processing pythonImage processing python
Image processing python
 
“Using an ISP for Real-time Data Augmentation,” a Presentation from Pony.AI
“Using an ISP for Real-time Data Augmentation,” a Presentation from Pony.AI“Using an ISP for Real-time Data Augmentation,” a Presentation from Pony.AI
“Using an ISP for Real-time Data Augmentation,” a Presentation from Pony.AI
 
final presentation
final presentationfinal presentation
final presentation
 
Tracking Robustness and Green View Index Estimation of Augmented and Diminish...
Tracking Robustness and Green View Index Estimation of Augmented and Diminish...Tracking Robustness and Green View Index Estimation of Augmented and Diminish...
Tracking Robustness and Green View Index Estimation of Augmented and Diminish...
 
Smart Control of Traffic Signal System using Image Processing
Smart Control of Traffic Signal System using Image Processing Smart Control of Traffic Signal System using Image Processing
Smart Control of Traffic Signal System using Image Processing
 
Applications of Image Processing and Real-Time embedded Systems in Autonomous...
Applications of Image Processing and Real-Time embedded Systems in Autonomous...Applications of Image Processing and Real-Time embedded Systems in Autonomous...
Applications of Image Processing and Real-Time embedded Systems in Autonomous...
 
TraVis CTTHES3
TraVis CTTHES3TraVis CTTHES3
TraVis CTTHES3
 
Quicklook technology assessment topmod software_ccchittim
Quicklook technology assessment topmod software_ccchittimQuicklook technology assessment topmod software_ccchittim
Quicklook technology assessment topmod software_ccchittim
 
Digital Image Forensics: camera fingerprint and its robustness
Digital Image Forensics: camera fingerprint and its robustness Digital Image Forensics: camera fingerprint and its robustness
Digital Image Forensics: camera fingerprint and its robustness
 
Panorama Technique for 3D Animation movie, Design and Evaluating
Panorama Technique for 3D Animation movie, Design and EvaluatingPanorama Technique for 3D Animation movie, Design and Evaluating
Panorama Technique for 3D Animation movie, Design and Evaluating
 
FINAL YEAR PROJECT1_3
FINAL YEAR PROJECT1_3FINAL YEAR PROJECT1_3
FINAL YEAR PROJECT1_3
 
IRJET- Reckoning the Vehicle using MATLAB
IRJET- Reckoning the Vehicle using MATLABIRJET- Reckoning the Vehicle using MATLAB
IRJET- Reckoning the Vehicle using MATLAB
 
Capturing Real-World Materials for Real-Time Development - Unite LA
Capturing Real-World Materials for Real-Time Development - Unite LACapturing Real-World Materials for Real-Time Development - Unite LA
Capturing Real-World Materials for Real-Time Development - Unite LA
 
Face in video evaluation (five)
Face in video evaluation (five)Face in video evaluation (five)
Face in video evaluation (five)
 
Gutter Detection and Ranging
Gutter Detection and RangingGutter Detection and Ranging
Gutter Detection and Ranging
 
Rapid Laser Scanning the process
Rapid Laser Scanning the processRapid Laser Scanning the process
Rapid Laser Scanning the process
 
Sample 2018 top 5 3 d time-of-flight image sensor players in north america, e...
Sample 2018 top 5 3 d time-of-flight image sensor players in north america, e...Sample 2018 top 5 3 d time-of-flight image sensor players in north america, e...
Sample 2018 top 5 3 d time-of-flight image sensor players in north america, e...
 
APPLICATION OF IP TECHNIQUES IN TRAFFIC CONTROL SYSTEM
APPLICATION OF IP TECHNIQUES IN TRAFFIC CONTROL SYSTEMAPPLICATION OF IP TECHNIQUES IN TRAFFIC CONTROL SYSTEM
APPLICATION OF IP TECHNIQUES IN TRAFFIC CONTROL SYSTEM
 
IRJET- Dynamic Traffic Management System
IRJET- Dynamic Traffic Management SystemIRJET- Dynamic Traffic Management System
IRJET- Dynamic Traffic Management System
 
Photosynth
PhotosynthPhotosynth
Photosynth
 

Recently uploaded

Capitol Tech U Doctoral Presentation - April 2024.pptx
Capitol Tech U Doctoral Presentation - April 2024.pptxCapitol Tech U Doctoral Presentation - April 2024.pptx
Capitol Tech U Doctoral Presentation - April 2024.pptxCapitolTechU
 
Solving Puzzles Benefits Everyone (English).pptx
Solving Puzzles Benefits Everyone (English).pptxSolving Puzzles Benefits Everyone (English).pptx
Solving Puzzles Benefits Everyone (English).pptxOH TEIK BIN
 
Meghan Sutherland In Media Res Media Component
Meghan Sutherland In Media Res Media ComponentMeghan Sutherland In Media Res Media Component
Meghan Sutherland In Media Res Media ComponentInMediaRes1
 
DATA STRUCTURE AND ALGORITHM for beginners
DATA STRUCTURE AND ALGORITHM for beginnersDATA STRUCTURE AND ALGORITHM for beginners
DATA STRUCTURE AND ALGORITHM for beginnersSabitha Banu
 
Introduction to AI in Higher Education_draft.pptx
Introduction to AI in Higher Education_draft.pptxIntroduction to AI in Higher Education_draft.pptx
Introduction to AI in Higher Education_draft.pptxpboyjonauth
 
Painted Grey Ware.pptx, PGW Culture of India
Painted Grey Ware.pptx, PGW Culture of IndiaPainted Grey Ware.pptx, PGW Culture of India
Painted Grey Ware.pptx, PGW Culture of IndiaVirag Sontakke
 
CARE OF CHILD IN INCUBATOR..........pptx
CARE OF CHILD IN INCUBATOR..........pptxCARE OF CHILD IN INCUBATOR..........pptx
CARE OF CHILD IN INCUBATOR..........pptxGaneshChakor2
 
Types of Journalistic Writing Grade 8.pptx
Types of Journalistic Writing Grade 8.pptxTypes of Journalistic Writing Grade 8.pptx
Types of Journalistic Writing Grade 8.pptxEyham Joco
 
How to Configure Email Server in Odoo 17
How to Configure Email Server in Odoo 17How to Configure Email Server in Odoo 17
How to Configure Email Server in Odoo 17Celine George
 
Employee wellbeing at the workplace.pptx
Employee wellbeing at the workplace.pptxEmployee wellbeing at the workplace.pptx
Employee wellbeing at the workplace.pptxNirmalaLoungPoorunde1
 
Proudly South Africa powerpoint Thorisha.pptx
Proudly South Africa powerpoint Thorisha.pptxProudly South Africa powerpoint Thorisha.pptx
Proudly South Africa powerpoint Thorisha.pptxthorishapillay1
 
“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...Marc Dusseiller Dusjagr
 
History Class XII Ch. 3 Kinship, Caste and Class (1).pptx
History Class XII Ch. 3 Kinship, Caste and Class (1).pptxHistory Class XII Ch. 3 Kinship, Caste and Class (1).pptx
History Class XII Ch. 3 Kinship, Caste and Class (1).pptxsocialsciencegdgrohi
 
Enzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdf
Enzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdfEnzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdf
Enzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdfSumit Tiwari
 
ECONOMIC CONTEXT - PAPER 1 Q3: NEWSPAPERS.pptx
ECONOMIC CONTEXT - PAPER 1 Q3: NEWSPAPERS.pptxECONOMIC CONTEXT - PAPER 1 Q3: NEWSPAPERS.pptx
ECONOMIC CONTEXT - PAPER 1 Q3: NEWSPAPERS.pptxiammrhaywood
 
call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️
call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️
call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️9953056974 Low Rate Call Girls In Saket, Delhi NCR
 

Recently uploaded (20)

Capitol Tech U Doctoral Presentation - April 2024.pptx
Capitol Tech U Doctoral Presentation - April 2024.pptxCapitol Tech U Doctoral Presentation - April 2024.pptx
Capitol Tech U Doctoral Presentation - April 2024.pptx
 
Solving Puzzles Benefits Everyone (English).pptx
Solving Puzzles Benefits Everyone (English).pptxSolving Puzzles Benefits Everyone (English).pptx
Solving Puzzles Benefits Everyone (English).pptx
 
Meghan Sutherland In Media Res Media Component
Meghan Sutherland In Media Res Media ComponentMeghan Sutherland In Media Res Media Component
Meghan Sutherland In Media Res Media Component
 
DATA STRUCTURE AND ALGORITHM for beginners
DATA STRUCTURE AND ALGORITHM for beginnersDATA STRUCTURE AND ALGORITHM for beginners
DATA STRUCTURE AND ALGORITHM for beginners
 
Introduction to AI in Higher Education_draft.pptx
Introduction to AI in Higher Education_draft.pptxIntroduction to AI in Higher Education_draft.pptx
Introduction to AI in Higher Education_draft.pptx
 
Painted Grey Ware.pptx, PGW Culture of India
Painted Grey Ware.pptx, PGW Culture of IndiaPainted Grey Ware.pptx, PGW Culture of India
Painted Grey Ware.pptx, PGW Culture of India
 
Model Call Girl in Bikash Puri Delhi reach out to us at 🔝9953056974🔝
Model Call Girl in Bikash Puri  Delhi reach out to us at 🔝9953056974🔝Model Call Girl in Bikash Puri  Delhi reach out to us at 🔝9953056974🔝
Model Call Girl in Bikash Puri Delhi reach out to us at 🔝9953056974🔝
 
CARE OF CHILD IN INCUBATOR..........pptx
CARE OF CHILD IN INCUBATOR..........pptxCARE OF CHILD IN INCUBATOR..........pptx
CARE OF CHILD IN INCUBATOR..........pptx
 
ESSENTIAL of (CS/IT/IS) class 06 (database)
ESSENTIAL of (CS/IT/IS) class 06 (database)ESSENTIAL of (CS/IT/IS) class 06 (database)
ESSENTIAL of (CS/IT/IS) class 06 (database)
 
Types of Journalistic Writing Grade 8.pptx
Types of Journalistic Writing Grade 8.pptxTypes of Journalistic Writing Grade 8.pptx
Types of Journalistic Writing Grade 8.pptx
 
How to Configure Email Server in Odoo 17
How to Configure Email Server in Odoo 17How to Configure Email Server in Odoo 17
How to Configure Email Server in Odoo 17
 
Employee wellbeing at the workplace.pptx
Employee wellbeing at the workplace.pptxEmployee wellbeing at the workplace.pptx
Employee wellbeing at the workplace.pptx
 
9953330565 Low Rate Call Girls In Rohini Delhi NCR
9953330565 Low Rate Call Girls In Rohini  Delhi NCR9953330565 Low Rate Call Girls In Rohini  Delhi NCR
9953330565 Low Rate Call Girls In Rohini Delhi NCR
 
Proudly South Africa powerpoint Thorisha.pptx
Proudly South Africa powerpoint Thorisha.pptxProudly South Africa powerpoint Thorisha.pptx
Proudly South Africa powerpoint Thorisha.pptx
 
“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
“Oh GOSH! Reflecting on Hackteria's Collaborative Practices in a Global Do-It...
 
History Class XII Ch. 3 Kinship, Caste and Class (1).pptx
History Class XII Ch. 3 Kinship, Caste and Class (1).pptxHistory Class XII Ch. 3 Kinship, Caste and Class (1).pptx
History Class XII Ch. 3 Kinship, Caste and Class (1).pptx
 
Enzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdf
Enzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdfEnzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdf
Enzyme, Pharmaceutical Aids, Miscellaneous Last Part of Chapter no 5th.pdf
 
ECONOMIC CONTEXT - PAPER 1 Q3: NEWSPAPERS.pptx
ECONOMIC CONTEXT - PAPER 1 Q3: NEWSPAPERS.pptxECONOMIC CONTEXT - PAPER 1 Q3: NEWSPAPERS.pptx
ECONOMIC CONTEXT - PAPER 1 Q3: NEWSPAPERS.pptx
 
call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️
call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️
call girls in Kamla Market (DELHI) 🔝 >༒9953330565🔝 genuine Escort Service 🔝✔️✔️
 
TataKelola dan KamSiber Kecerdasan Buatan v022.pdf
TataKelola dan KamSiber Kecerdasan Buatan v022.pdfTataKelola dan KamSiber Kecerdasan Buatan v022.pdf
TataKelola dan KamSiber Kecerdasan Buatan v022.pdf
 

Automatic Photo Selection For Media And Entertainment Applications

  • 1. Automatic Photo Selection for Media and Entertainment Applications Ekaterina Potapova, Marta Egorova, Ilia Safonov National Nuclear Research University MEPhI Moscow, Russia GraphiCon 2009 5-9 October
  • 2. Applications Automatic Photo Selection for Media and Entertainment Applications GraphiCon 2009 2
  • 3. Applications Automatic Photo Selection for Media and Entertainment Applications GraphiCon 2009 2
  • 4. Applications – photo book Images are taken from printbook.ru, ehow.com, snapfish.com.au, smilebooks.co.uk GraphiCon 2009 3 Automatic Photo Selection for Media and Entertainment Applications
  • 5. Applications – slide show Photos from ITaS’2008 GraphiCon 2009 4 Automatic Photo Selection for Media and Entertainment Applications
  • 6. General workflow GraphiCon 2009 5 Automatic Photo Selection for Media and Entertainment Applications
  • 7. GraphiCon 2009 5 Automatic Photo Selection for Media and Entertainment Applications General workflow Detection of low-quality photos
  • 8. General workflow GraphiCon 2009 5 Automatic Photo Selection for Media and Entertainment Applications Detection of low-quality photos Adaptive quantization on time-camera plane
  • 9. General workflow GraphiCon 2009 5 Automatic Photo Selection for Media and Entertainment Applications Selection of appealing photos Detection of low-quality photos Adaptive quantization on time-camera plane
  • 10. Detection of low-quality photos GraphiCon 2009 6 Automatic Photo Selection for Media and Entertainment Applications
  • 11. Estimation of JPEG quality A.Foi et al.,2007 Images are taken from en.wikipedia.org Quantization Table GraphiCon 2009 7 Automatic Photo Selection for Media and Entertainment Applications
  • 12. Detection of backlit, low-contrast & blurred photos Two Ada Boost classifiers committee: -for detection of low-contrast and backlit photos -for detection of blurred photos GraphiCon 2009 8 Automatic Photo Selection for Media and Entertainment Applications + Good photo Bad photo True False … …
  • 13. Detection of backlit and low-contrast photos - 1 S1/S2 - ratio of tones in shadows to midtones GraphiCon 2009 9 Automatic Photo Selection for Media and Entertainment Applications
  • 14. S11/S12 - ratio of tones in first to second part of shadows Detection of backlit and low-contrast photos - 1 GraphiCon 2009 9 Automatic Photo Selection for Media and Entertainment Applications
  • 15. M1/M2 - ratio of the histogram maximum in shadows to the maximum in midtones Detection of backlit and low-contrast photos - 1 GraphiCon 2009 9 Automatic Photo Selection for Media and Entertainment Applications
  • 16. P1 - location of the histogram maximum in shadows P1 Detection of backlit and low-contrast photos - 1 GraphiCon 2009 9 Automatic Photo Selection for Media and Entertainment Applications
  • 17. C – global contrast H 0 C 0 C 1 H 1 Detection of backlit and low-contrast photos - 1 GraphiCon 2009 9 Automatic Photo Selection for Media and Entertainment Applications
  • 18. Training set: 480 photos Error rate on cross-validation test : ~0.055 Testing set: 1830 with 2% affected by backlit and low-contrast photos The number of False Positives (FP) is 10 The number of False Negatives (FN) is 3 Low-contrast photo Backlit photo Detection of backlit and low-contrast photos - 2 GraphiCon 2009 10 Automatic Photo Selection for Media and Entertainment Applications
  • 19. Image Intensity image Z 1 =[-1 1] Z 2 =[-1 0 1] Z 3 =[-1 0 0 1] Z 10 =[-1 0 0 0 0 0 0 0 0 0 1] I.Safonov et al.,2008 … Edge image Histogram Normalized entropy Entropy to [0, 1] ? ? ? ? An An GraphiCon 2009 11 Detection of blurred photos Automatic Photo Selection for Media and Entertainment Applications
  • 20. Crete et al., 2007 F.Crete et al.,2007 ? Image Blurred image Edge image Edge image Comparison of the images HPF=[1 -1] LPF=[1 1 1 1 1 1 1 1 1]/9 Detection of blurred photos GraphiCon 2009 11 Automatic Photo Selection for Media and Entertainment Applications
  • 21. Training set: 416 photos Error rate on cross-validation test : ~0.07 Testing set: 1830 with 171 blurred photos The number of False Positives (FP) is 34 The number of False Negatives (FN) is 10 Detection of blurred photos GraphiCon 2009 11 Automatic Photo Selection for Media and Entertainment Applications
  • 22. Time and camera-based quantization i is an index of source L is time between the least and the most time for the largest source Nps is a number of sources H = L/M M is count of images GraphiCon 2009 11 Automatic Photo Selection for Media and Entertainment Applications N region < N group N region < M Calculation of bounding boxes Partition into 2 app. equal subregions Seeking for the biggest region 1200 3600 2400 7200 0 36000 T, s 21600
  • 23. GraphiCon 2009 12 Automatic Photo Selection for Media and Entertainment Applications Salient Photo Selection The most appealing photo is the most salient photo L.Itti, C.Koch et al. Images are taken from the Internet
  • 24. Conspicuity maps Gaussian pyramids Image Intensity image r-channel g-channel b-channel R-channel G-channel B-channel Y-channel Orientation map Intensity map Color map Saliency map Feature maps Gabor pyramids GraphiCon 2009 13 Automatic Photo Selection for Media and Entertainment Applications Salient Photo Selection
  • 25. original image saliency map intensity map color map orientation map ROI Automatic Photo Selection for Media and Entertainment Applications Salient Photo Selection GraphiCon 2009 14 Image is taken from the Internet
  • 26. Automatic Photo Selection for Media and Entertainment Applications Salient Photo Selection GraphiCon 2009 15 124 88 11 100 81 92 62 83 105 70 Saliency Index
  • 27. Automatic Photo Selection for Media and Entertainment Applications Salient Photo Selection GraphiCon 2009 15 83 11 124 Saliency Index 81 88 62 92 105 70 100
  • 28.
  • 29. Photos ranking Heuristic formula, experiments have shown that value w=25 gives the best result Automatic Photo Selection for Media and Entertainment Applications GraphiCon 2009 17 124 88 11 116 92 118 148 95 62 100
  • 30. Photos ranking Heuristic formula, experiments have shown that value w=25 gives the best result Automatic Photo Selection for Media and Entertainment Applications GraphiCon 2009 17 118 62 124 88 11 100 116 92 148 95
  • 31. Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 18
  • 32. Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 18 Autocollage choice Our choice
  • 33. Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 18
  • 34. Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 18 Autocollage choice Our choice
  • 35. Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 18
  • 36. Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 18 Autocollage choice Our choice
  • 37. Automatic Photo Selection for Media and Entertainment Applications Results and discussion GraphiCon 2009 19 Proposed AutoCollage Random 14 1 4 3 3 3 Unacceptable 21 5 2 4 5 5 Acceptable 15 4 4 3 2 2 Agree with experts 9 1 4 1 1 2 Unacceptable 24 4 0 7 7 6 Acceptable 17 5 6 2 2 2 Agree with experts 4 1 1 0 1 1 Unacceptable 17 2 4 4 4 3 Acceptable 29 7 5 6 5 6 Agree with experts Sum Set 5 Set 4 Set 3 Set 2 Set 1
  • 38. ? Automatic Photo Selection for Media and Entertainment Applications Questions & Answers GraphiCon 2009 8
  • 39. Automatic Photo Selection for Media and Entertainment Applications GraphiCon 2009 9 Thank you for your attention =)