SlideShare une entreprise Scribd logo
1  sur  1
Leveraging an Image Folksonomy and the Signature Quadratic Form
     Distance for Semantic-Based Detection of Near-Duplicate Video Clips
                                                    Hyun-seok Min, Jae Young Choi, Wesley De Neve, and Yong Man Ro
                                                                                         Image and Video Systems Lab
                                                                            Korea Advanced Institute of Science and Technology (KAIST)
                                                                                              Daejeon, South Korea
                                                e-mail: hsmin@kaist.ac.kr                                                                                                         website: http://ivylab.kaist.ac.kr

I. INTRODUCTION                                                                                                                                     IV. EXPERIMENTS
- Observations                                                                                                                                      1. Experimental setup
   - an increasing number of near-duplicate video clips (NDVCs) can be                                                                              - Use of TRECVID 2009 for creating NDVCs and reference video clips
     found on websites for video sharing                                                                                                            - Use of MIRFLICKR-25000 as a source of collective knowledge
   - content transformations tend to preserve semantic information                                                                                  - Use of VIREO-374 for model-based semantic concept detection
- Novel idea
   - NDVC detection using semantic concept detection                                                                                                2. Experimental results
- Research challenges                                                                                                                               2.1. Influence of semantic concept popularity
   - semantic coverage: use of model-free semantic concept detection                                                                                  - The effectiveness of model-based semantic concept detection highly
   - semantic similarity: use of adaptive semantic distance measurement                                                                                 depends on the popularity of the semantic concept models used
II. SEMANTIC VIDEO SIGNATURE CREATION USING AN                                                                                                           - non-popular semantic concept models hardly contribute to
    IMAGE FOLKSONOMY                                                                                                                                       improving the effectiveness of NDVC detection
                                                                                                                                                             1.2

                                                                                                                                                              1
                 Input shot Si
                                                                                                                                                             0.8




                                                                                                                                                      NDCR
                                                                       Visual                        Image folksonomy F
                                                                                                                                                             0.6
         Extraction of low-level visual features                                                                                                                                              Descending order of popularity
                                                                      features                                             User
                                                                                                                                                             0.4                              Ascending order of popularity
                                                                                                    User
              Content-based image retrieval                                                                     User-contributed images
                                                                                                                  User-supplied tags                         0.2
                                                                      Images         User-contributed images
             k nearest visual neighbors of Si                         & tags            User-supplied tags
                                                                                                                                                              0




                                                                                                                                                                   120




                                                                                                                                                                   310
                                                                                                                                                                    10
                                                                                                                                                                    20
                                                                                                                                                                    30
                                                                                                                                                                    40
                                                                                                                                                                    50
                                                                                                                                                                    60
                                                                                                                                                                    70
                                                                                                                                                                    80
                                                                                                                                                                    90
                                                                                                                                                                   100
                                                                                                                                                                   110

                                                                                                                                                                   130
                                                                                                                                                                   140
                                                                                                                                                                   150
                                                                                                                                                                   160
                                                                                                                                                                   170
                                                                                                                                                                   180
                                                                                                                                                                   190
                                                                                                                                                                   200
                                                                                                                                                                   210
                                                                                                                                                                   220
                                                                                                                                                                   230
                                                                                                                                                                   240
                                                                                                                                                                   250
                                                                                                                                                                   260
                                                                                                                                                                   270
                                                                                                                                                                   280
                                                                                                                                                                   290
                                                                                                                                                                   300

                                                                                                                                                                   320
                                                                                                                                                                   330
                                                                                                                                                                   340
                                                                                                                                                                   350
                                                                                                                                                                   360
                                                                                                                                                                   370
                  : night, sky, stars, mountains, milkyway, aquila,                                             User-contributed images
                  sagittarius, scorpius, ...                                         User-contributed images                                                                                     Number of semantic concepts used
                                                                                                                   User-supplied tags
                                                                                       User-supplied tags
                  : milkyway, sky, space, astrophotography,
                                                                                                                                                                   Fig. 2. Influence of semantic concept popularity on NDVC detection.
                  night, telescope, jupiter, clouds, ...
                                                                                                    User
                                                                                                                           User                     2.2. Influence of different types of video content
       ...




                                                                                                                                                      - To facilitate effective NDVC detection, video signatures need to be
                  : milky way, galaxy, stars, sky                                                                                                       robust against the use of different types of video content
                                                                                                                                                         - category 1 (documentaries), category 2 (news),
                                                                                                                                                           category 3 (drama and movies), category 4 (miscellaneous)
Fig. 1. Retrieval of the k nearest visual neighbor images and their associated tags
                   from an image folksonomy F for a video shot Si.
                                                                                                                                                      - The effectiveness of the proposed NDVC detection technique is
                                                                                                                                                        stable and high for all types of video content investigated
- Metric for measuring the relevance of a tag t w.r.t. the shot Si:
                                                c   : the frequency of t in the set of k neighbors
          c Lt
  R (t ) = -   , Lt                                 : the number of images labeled with t in F
          K F
                 F                                   : the number of images in F
- Layout of the semantic feature signature Ai of a shot Si:

             [                                                    ]
Ai = ti , j , wi , j , j = 1,..., Ai , wi , j : a weight value for tag ti,j

- Computation of the weight value for tag ti,j :                                                                    R(ti , j )
                                                                                                wi , j =         Ai                                          Fig. 3. Effectiveness of NDVC detection for different types of video content.

                                                                                                                ∑ R(ti, k )                                    Key frame
                                                                                                                                                                             Model-based
                                                                                                                                                                              approach
                                                                                                                                                                                              Model-free
                                                                                                                                                                                              approach
                                                                                                                                                                                                                 Key frame
                                                                                                                                                                                                                                    Model-based
                                                                                                                                                                                                                                     approach
                                                                                                                                                                                                                                                   Model-free
                                                                                                                                                                                                                                                   approach
                                                                                                                k =1                                                                           Cloud                                                 Stars
                                                                                                                                                                                                Sky                                                  Night
                                                                                                                                                                                               Water                                               Geotagged
                                                                                                                                                                                 N/A                                                   N/A
III. SEMANTIC DISTANCE MEASUREMENT USING THE                                                                                                                                                  Moonlight
                                                                                                                                                                                              Rainbow
                                                                                                                                                                                                                                                  Constellation
                                                                                                                                                                                                                                                      Sky
     SIGNATURE QUADRATIC FORM DISTANCE (SQFD)                                                                                                                                                    …                                                    …

- Adaptive semantic distance measurement between shots Sq and Sr:                                                                                                                               She
                                                                                                                                                                                                                                                     Puppy
                                                                                                                                                                                                                                                      Dog
                                                                                                                                          r T
                                                                                                                                                                                                Blue
                                                                                        w |- w G w |- w
                       q            r                        q          r                   q               r             q
 Dshot (S , S ) = SQFD(A , A ) =
                                                                                                                                                                            Civilian Person                                                          Grass
                                                                                                                                                ,                               Group
                                                                                                                                                                                               Clouds
                                                                                                                                                                                                Zoo
                                                                                                                                                                                                                                       N/A
                                                                                                                                                                                                                                                    Summer
                                                                                                                                                                                                                                                     Safari
                                                                                                                                                                                                 …
                                                                                                                                                                                                                                                       …
     q                         q          q                   r                  r              r
 w                 w ,...,w    1
                                          Aq
                                                         w                  w ,...,w
                                                                                 1
                                                                                                                                                                       Fig. 4. Example key frames with detected semantic concepts
                                                                                                Ar                                                                    (underlined semantic concepts are considered to be correct).
                                                                                                                                                    V. CONCLUSIONS
- The elements of the ground similarity matrix G:                                                                                                   -This paper discussed a novel technique for NDVC detection
                                                                                                                                                        - takes advantage of the collective knowledge in an image folksonomy
              It
                   i           tj       I ti ∩ t j : the set of images annotated with both tag ti and tj                                                   - allows using an unrestricted and dynamic concept vocabulary
 gij                                ,                                                                                                                   - takes advantage of the flexible SQFD metric
                  It                     I ti   : the set of images annotated with tag ti                                                                  - allows taking into account that the nature, the relevance, and the
                           i                                                                                                                                 number of semantic concepts may strongly vary from shot to shot

                                                    IEEE International Conference on Multimedia and Expo (ICME), July 2011, Barcelona (Spain)

Contenu connexe

Similaire à Leveraging an image folksonomy and the signature quadratic form distance for semantic based detection of near-duplicate video clips

Towards a Better Understanding of Model-Free Semantic Concept Detection for A...
Towards a Better Understanding of Model-Free Semantic Concept Detection for A...Towards a Better Understanding of Model-Free Semantic Concept Detection for A...
Towards a Better Understanding of Model-Free Semantic Concept Detection for A...Wesley De Neve
 
Content based image retrieval Projects.pdf
Content based image retrieval Projects.pdfContent based image retrieval Projects.pdf
Content based image retrieval Projects.pdfrupaymts
 
GeniUS: Generic User Modeling Library for the Social Semantic Web
GeniUS: Generic User Modeling Library for the Social Semantic WebGeniUS: Generic User Modeling Library for the Social Semantic Web
GeniUS: Generic User Modeling Library for the Social Semantic WebWeb Information Systems, TU Delft
 
Elettronica: Multimedia Information Processing in Smart Environments by Aless...
Elettronica: Multimedia Information Processing in Smart Environments by Aless...Elettronica: Multimedia Information Processing in Smart Environments by Aless...
Elettronica: Multimedia Information Processing in Smart Environments by Aless...Codemotion
 

Similaire à Leveraging an image folksonomy and the signature quadratic form distance for semantic based detection of near-duplicate video clips (7)

Towards a Better Understanding of Model-Free Semantic Concept Detection for A...
Towards a Better Understanding of Model-Free Semantic Concept Detection for A...Towards a Better Understanding of Model-Free Semantic Concept Detection for A...
Towards a Better Understanding of Model-Free Semantic Concept Detection for A...
 
Dohyoung lee icassp2012_poster
Dohyoung lee icassp2012_posterDohyoung lee icassp2012_poster
Dohyoung lee icassp2012_poster
 
L01web 2x2
L01web 2x2L01web 2x2
L01web 2x2
 
ORUSSI: Optimal Road sUrveillance based on Scalable vIdeo
ORUSSI: Optimal Road sUrveillance based on Scalable vIdeoORUSSI: Optimal Road sUrveillance based on Scalable vIdeo
ORUSSI: Optimal Road sUrveillance based on Scalable vIdeo
 
Content based image retrieval Projects.pdf
Content based image retrieval Projects.pdfContent based image retrieval Projects.pdf
Content based image retrieval Projects.pdf
 
GeniUS: Generic User Modeling Library for the Social Semantic Web
GeniUS: Generic User Modeling Library for the Social Semantic WebGeniUS: Generic User Modeling Library for the Social Semantic Web
GeniUS: Generic User Modeling Library for the Social Semantic Web
 
Elettronica: Multimedia Information Processing in Smart Environments by Aless...
Elettronica: Multimedia Information Processing in Smart Environments by Aless...Elettronica: Multimedia Information Processing in Smart Environments by Aless...
Elettronica: Multimedia Information Processing in Smart Environments by Aless...
 

Plus de Wesley De Neve

Towards diagnosis of rotator cuff tears in 3-D MRI using 3-D convolutional ne...
Towards diagnosis of rotator cuff tears in 3-D MRI using 3-D convolutional ne...Towards diagnosis of rotator cuff tears in 3-D MRI using 3-D convolutional ne...
Towards diagnosis of rotator cuff tears in 3-D MRI using 3-D convolutional ne...Wesley De Neve
 
Investigating the biological relevance in trained embedding representations o...
Investigating the biological relevance in trained embedding representations o...Investigating the biological relevance in trained embedding representations o...
Investigating the biological relevance in trained embedding representations o...Wesley De Neve
 
Impact of adversarial examples on deep learning models for biomedical image s...
Impact of adversarial examples on deep learning models for biomedical image s...Impact of adversarial examples on deep learning models for biomedical image s...
Impact of adversarial examples on deep learning models for biomedical image s...Wesley De Neve
 
Learning Biologically Relevant Features Using Convolutional Neural Networks f...
Learning Biologically Relevant Features Using Convolutional Neural Networks f...Learning Biologically Relevant Features Using Convolutional Neural Networks f...
Learning Biologically Relevant Features Using Convolutional Neural Networks f...Wesley De Neve
 
The 5th Aslla Symposium
The 5th Aslla SymposiumThe 5th Aslla Symposium
The 5th Aslla SymposiumWesley De Neve
 
Ghent University Global Campus 101
Ghent University Global Campus 101Ghent University Global Campus 101
Ghent University Global Campus 101Wesley De Neve
 
Booklet for the First GUGC Research Symposium
Booklet for the First GUGC Research SymposiumBooklet for the First GUGC Research Symposium
Booklet for the First GUGC Research SymposiumWesley De Neve
 
Center for Biotech Data Science at Ghent University Global Campus
Center for Biotech Data Science at Ghent University Global CampusCenter for Biotech Data Science at Ghent University Global Campus
Center for Biotech Data Science at Ghent University Global CampusWesley De Neve
 
Center for Biotech Data Science at Ghent University Global Campus
Center for Biotech Data Science at Ghent University Global CampusCenter for Biotech Data Science at Ghent University Global Campus
Center for Biotech Data Science at Ghent University Global CampusWesley De Neve
 
Learning biologically relevant features using convolutional neural networks f...
Learning biologically relevant features using convolutional neural networks f...Learning biologically relevant features using convolutional neural networks f...
Learning biologically relevant features using convolutional neural networks f...Wesley De Neve
 
Towards reading genomic data using deep learning-driven NLP techniques
Towards reading genomic data using deep learning-driven NLP techniquesTowards reading genomic data using deep learning-driven NLP techniques
Towards reading genomic data using deep learning-driven NLP techniquesWesley De Neve
 
Deep Machine Learning for Making Sense of Biotech Data - From Clean Energy to...
Deep Machine Learning for Making Sense of Biotech Data - From Clean Energy to...Deep Machine Learning for Making Sense of Biotech Data - From Clean Energy to...
Deep Machine Learning for Making Sense of Biotech Data - From Clean Energy to...Wesley De Neve
 
GUGC Info Session - Informatics and Bioinformatics
GUGC Info Session - Informatics and BioinformaticsGUGC Info Session - Informatics and Bioinformatics
GUGC Info Session - Informatics and BioinformaticsWesley De Neve
 
Ghent University Global Campus - Sungkyunkwan University: Workshop on Researc...
Ghent University Global Campus - Sungkyunkwan University: Workshop on Researc...Ghent University Global Campus - Sungkyunkwan University: Workshop on Researc...
Ghent University Global Campus - Sungkyunkwan University: Workshop on Researc...Wesley De Neve
 
Ghent University and GUGC-K: Overview of Teaching and Research Activities
Ghent University and GUGC-K: Overview of Teaching and Research ActivitiesGhent University and GUGC-K: Overview of Teaching and Research Activities
Ghent University and GUGC-K: Overview of Teaching and Research ActivitiesWesley De Neve
 
Biotech Data Science @ GUGC in Korea: Deep Learning for Prediction of Drug-Ta...
Biotech Data Science @ GUGC in Korea: Deep Learning for Prediction of Drug-Ta...Biotech Data Science @ GUGC in Korea: Deep Learning for Prediction of Drug-Ta...
Biotech Data Science @ GUGC in Korea: Deep Learning for Prediction of Drug-Ta...Wesley De Neve
 
Exploring Deep Machine Learning for Automatic Right Whale Recognition and No...
 Exploring Deep Machine Learning for Automatic Right Whale Recognition and No... Exploring Deep Machine Learning for Automatic Right Whale Recognition and No...
Exploring Deep Machine Learning for Automatic Right Whale Recognition and No...Wesley De Neve
 
Deep Machine Learning for Automating Biotech Tasks Through Self-Learning Expe...
Deep Machine Learning for Automating Biotech Tasks Through Self-Learning Expe...Deep Machine Learning for Automating Biotech Tasks Through Self-Learning Expe...
Deep Machine Learning for Automating Biotech Tasks Through Self-Learning Expe...Wesley De Neve
 
Towards using multimedia technology for biological data processing
Towards using multimedia technology for biological data processingTowards using multimedia technology for biological data processing
Towards using multimedia technology for biological data processingWesley De Neve
 
Multimedia Lab @ Ghent University - iMinds - Organizational Overview & Outlin...
Multimedia Lab @ Ghent University - iMinds - Organizational Overview & Outlin...Multimedia Lab @ Ghent University - iMinds - Organizational Overview & Outlin...
Multimedia Lab @ Ghent University - iMinds - Organizational Overview & Outlin...Wesley De Neve
 

Plus de Wesley De Neve (20)

Towards diagnosis of rotator cuff tears in 3-D MRI using 3-D convolutional ne...
Towards diagnosis of rotator cuff tears in 3-D MRI using 3-D convolutional ne...Towards diagnosis of rotator cuff tears in 3-D MRI using 3-D convolutional ne...
Towards diagnosis of rotator cuff tears in 3-D MRI using 3-D convolutional ne...
 
Investigating the biological relevance in trained embedding representations o...
Investigating the biological relevance in trained embedding representations o...Investigating the biological relevance in trained embedding representations o...
Investigating the biological relevance in trained embedding representations o...
 
Impact of adversarial examples on deep learning models for biomedical image s...
Impact of adversarial examples on deep learning models for biomedical image s...Impact of adversarial examples on deep learning models for biomedical image s...
Impact of adversarial examples on deep learning models for biomedical image s...
 
Learning Biologically Relevant Features Using Convolutional Neural Networks f...
Learning Biologically Relevant Features Using Convolutional Neural Networks f...Learning Biologically Relevant Features Using Convolutional Neural Networks f...
Learning Biologically Relevant Features Using Convolutional Neural Networks f...
 
The 5th Aslla Symposium
The 5th Aslla SymposiumThe 5th Aslla Symposium
The 5th Aslla Symposium
 
Ghent University Global Campus 101
Ghent University Global Campus 101Ghent University Global Campus 101
Ghent University Global Campus 101
 
Booklet for the First GUGC Research Symposium
Booklet for the First GUGC Research SymposiumBooklet for the First GUGC Research Symposium
Booklet for the First GUGC Research Symposium
 
Center for Biotech Data Science at Ghent University Global Campus
Center for Biotech Data Science at Ghent University Global CampusCenter for Biotech Data Science at Ghent University Global Campus
Center for Biotech Data Science at Ghent University Global Campus
 
Center for Biotech Data Science at Ghent University Global Campus
Center for Biotech Data Science at Ghent University Global CampusCenter for Biotech Data Science at Ghent University Global Campus
Center for Biotech Data Science at Ghent University Global Campus
 
Learning biologically relevant features using convolutional neural networks f...
Learning biologically relevant features using convolutional neural networks f...Learning biologically relevant features using convolutional neural networks f...
Learning biologically relevant features using convolutional neural networks f...
 
Towards reading genomic data using deep learning-driven NLP techniques
Towards reading genomic data using deep learning-driven NLP techniquesTowards reading genomic data using deep learning-driven NLP techniques
Towards reading genomic data using deep learning-driven NLP techniques
 
Deep Machine Learning for Making Sense of Biotech Data - From Clean Energy to...
Deep Machine Learning for Making Sense of Biotech Data - From Clean Energy to...Deep Machine Learning for Making Sense of Biotech Data - From Clean Energy to...
Deep Machine Learning for Making Sense of Biotech Data - From Clean Energy to...
 
GUGC Info Session - Informatics and Bioinformatics
GUGC Info Session - Informatics and BioinformaticsGUGC Info Session - Informatics and Bioinformatics
GUGC Info Session - Informatics and Bioinformatics
 
Ghent University Global Campus - Sungkyunkwan University: Workshop on Researc...
Ghent University Global Campus - Sungkyunkwan University: Workshop on Researc...Ghent University Global Campus - Sungkyunkwan University: Workshop on Researc...
Ghent University Global Campus - Sungkyunkwan University: Workshop on Researc...
 
Ghent University and GUGC-K: Overview of Teaching and Research Activities
Ghent University and GUGC-K: Overview of Teaching and Research ActivitiesGhent University and GUGC-K: Overview of Teaching and Research Activities
Ghent University and GUGC-K: Overview of Teaching and Research Activities
 
Biotech Data Science @ GUGC in Korea: Deep Learning for Prediction of Drug-Ta...
Biotech Data Science @ GUGC in Korea: Deep Learning for Prediction of Drug-Ta...Biotech Data Science @ GUGC in Korea: Deep Learning for Prediction of Drug-Ta...
Biotech Data Science @ GUGC in Korea: Deep Learning for Prediction of Drug-Ta...
 
Exploring Deep Machine Learning for Automatic Right Whale Recognition and No...
 Exploring Deep Machine Learning for Automatic Right Whale Recognition and No... Exploring Deep Machine Learning for Automatic Right Whale Recognition and No...
Exploring Deep Machine Learning for Automatic Right Whale Recognition and No...
 
Deep Machine Learning for Automating Biotech Tasks Through Self-Learning Expe...
Deep Machine Learning for Automating Biotech Tasks Through Self-Learning Expe...Deep Machine Learning for Automating Biotech Tasks Through Self-Learning Expe...
Deep Machine Learning for Automating Biotech Tasks Through Self-Learning Expe...
 
Towards using multimedia technology for biological data processing
Towards using multimedia technology for biological data processingTowards using multimedia technology for biological data processing
Towards using multimedia technology for biological data processing
 
Multimedia Lab @ Ghent University - iMinds - Organizational Overview & Outlin...
Multimedia Lab @ Ghent University - iMinds - Organizational Overview & Outlin...Multimedia Lab @ Ghent University - iMinds - Organizational Overview & Outlin...
Multimedia Lab @ Ghent University - iMinds - Organizational Overview & Outlin...
 

Dernier

"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii SoldatenkoFwdays
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfAlex Barbosa Coqueiro
 
Take control of your SAP testing with UiPath Test Suite
Take control of your SAP testing with UiPath Test SuiteTake control of your SAP testing with UiPath Test Suite
Take control of your SAP testing with UiPath Test SuiteDianaGray10
 
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptxThe Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptxLoriGlavin3
 
How AI, OpenAI, and ChatGPT impact business and software.
How AI, OpenAI, and ChatGPT impact business and software.How AI, OpenAI, and ChatGPT impact business and software.
How AI, OpenAI, and ChatGPT impact business and software.Curtis Poe
 
WordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your BrandWordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your Brandgvaughan
 
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024BookNet Canada
 
Hyperautomation and AI/ML: A Strategy for Digital Transformation Success.pdf
Hyperautomation and AI/ML: A Strategy for Digital Transformation Success.pdfHyperautomation and AI/ML: A Strategy for Digital Transformation Success.pdf
Hyperautomation and AI/ML: A Strategy for Digital Transformation Success.pdfPrecisely
 
Streamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupStreamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupFlorian Wilhelm
 
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek SchlawackFwdays
 
Digital Identity is Under Attack: FIDO Paris Seminar.pptx
Digital Identity is Under Attack: FIDO Paris Seminar.pptxDigital Identity is Under Attack: FIDO Paris Seminar.pptx
Digital Identity is Under Attack: FIDO Paris Seminar.pptxLoriGlavin3
 
TeamStation AI System Report LATAM IT Salaries 2024
TeamStation AI System Report LATAM IT Salaries 2024TeamStation AI System Report LATAM IT Salaries 2024
TeamStation AI System Report LATAM IT Salaries 2024Lonnie McRorey
 
unit 4 immunoblotting technique complete.pptx
unit 4 immunoblotting technique complete.pptxunit 4 immunoblotting technique complete.pptx
unit 4 immunoblotting technique complete.pptxBkGupta21
 
Advanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionAdvanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionDilum Bandara
 
Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Mattias Andersson
 
DevoxxFR 2024 Reproducible Builds with Apache Maven
DevoxxFR 2024 Reproducible Builds with Apache MavenDevoxxFR 2024 Reproducible Builds with Apache Maven
DevoxxFR 2024 Reproducible Builds with Apache MavenHervé Boutemy
 
Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 365Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 3652toLead Limited
 
DSPy a system for AI to Write Prompts and Do Fine Tuning
DSPy a system for AI to Write Prompts and Do Fine TuningDSPy a system for AI to Write Prompts and Do Fine Tuning
DSPy a system for AI to Write Prompts and Do Fine TuningLars Bell
 
A Deep Dive on Passkeys: FIDO Paris Seminar.pptx
A Deep Dive on Passkeys: FIDO Paris Seminar.pptxA Deep Dive on Passkeys: FIDO Paris Seminar.pptx
A Deep Dive on Passkeys: FIDO Paris Seminar.pptxLoriGlavin3
 
Generative AI for Technical Writer or Information Developers
Generative AI for Technical Writer or Information DevelopersGenerative AI for Technical Writer or Information Developers
Generative AI for Technical Writer or Information DevelopersRaghuram Pandurangan
 

Dernier (20)

"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdf
 
Take control of your SAP testing with UiPath Test Suite
Take control of your SAP testing with UiPath Test SuiteTake control of your SAP testing with UiPath Test Suite
Take control of your SAP testing with UiPath Test Suite
 
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptxThe Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
 
How AI, OpenAI, and ChatGPT impact business and software.
How AI, OpenAI, and ChatGPT impact business and software.How AI, OpenAI, and ChatGPT impact business and software.
How AI, OpenAI, and ChatGPT impact business and software.
 
WordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your BrandWordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your Brand
 
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
 
Hyperautomation and AI/ML: A Strategy for Digital Transformation Success.pdf
Hyperautomation and AI/ML: A Strategy for Digital Transformation Success.pdfHyperautomation and AI/ML: A Strategy for Digital Transformation Success.pdf
Hyperautomation and AI/ML: A Strategy for Digital Transformation Success.pdf
 
Streamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupStreamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project Setup
 
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
 
Digital Identity is Under Attack: FIDO Paris Seminar.pptx
Digital Identity is Under Attack: FIDO Paris Seminar.pptxDigital Identity is Under Attack: FIDO Paris Seminar.pptx
Digital Identity is Under Attack: FIDO Paris Seminar.pptx
 
TeamStation AI System Report LATAM IT Salaries 2024
TeamStation AI System Report LATAM IT Salaries 2024TeamStation AI System Report LATAM IT Salaries 2024
TeamStation AI System Report LATAM IT Salaries 2024
 
unit 4 immunoblotting technique complete.pptx
unit 4 immunoblotting technique complete.pptxunit 4 immunoblotting technique complete.pptx
unit 4 immunoblotting technique complete.pptx
 
Advanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionAdvanced Computer Architecture – An Introduction
Advanced Computer Architecture – An Introduction
 
Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?
 
DevoxxFR 2024 Reproducible Builds with Apache Maven
DevoxxFR 2024 Reproducible Builds with Apache MavenDevoxxFR 2024 Reproducible Builds with Apache Maven
DevoxxFR 2024 Reproducible Builds with Apache Maven
 
Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 365Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 365
 
DSPy a system for AI to Write Prompts and Do Fine Tuning
DSPy a system for AI to Write Prompts and Do Fine TuningDSPy a system for AI to Write Prompts and Do Fine Tuning
DSPy a system for AI to Write Prompts and Do Fine Tuning
 
A Deep Dive on Passkeys: FIDO Paris Seminar.pptx
A Deep Dive on Passkeys: FIDO Paris Seminar.pptxA Deep Dive on Passkeys: FIDO Paris Seminar.pptx
A Deep Dive on Passkeys: FIDO Paris Seminar.pptx
 
Generative AI for Technical Writer or Information Developers
Generative AI for Technical Writer or Information DevelopersGenerative AI for Technical Writer or Information Developers
Generative AI for Technical Writer or Information Developers
 

Leveraging an image folksonomy and the signature quadratic form distance for semantic based detection of near-duplicate video clips

  • 1. Leveraging an Image Folksonomy and the Signature Quadratic Form Distance for Semantic-Based Detection of Near-Duplicate Video Clips Hyun-seok Min, Jae Young Choi, Wesley De Neve, and Yong Man Ro Image and Video Systems Lab Korea Advanced Institute of Science and Technology (KAIST) Daejeon, South Korea e-mail: hsmin@kaist.ac.kr website: http://ivylab.kaist.ac.kr I. INTRODUCTION IV. EXPERIMENTS - Observations 1. Experimental setup - an increasing number of near-duplicate video clips (NDVCs) can be - Use of TRECVID 2009 for creating NDVCs and reference video clips found on websites for video sharing - Use of MIRFLICKR-25000 as a source of collective knowledge - content transformations tend to preserve semantic information - Use of VIREO-374 for model-based semantic concept detection - Novel idea - NDVC detection using semantic concept detection 2. Experimental results - Research challenges 2.1. Influence of semantic concept popularity - semantic coverage: use of model-free semantic concept detection - The effectiveness of model-based semantic concept detection highly - semantic similarity: use of adaptive semantic distance measurement depends on the popularity of the semantic concept models used II. SEMANTIC VIDEO SIGNATURE CREATION USING AN - non-popular semantic concept models hardly contribute to IMAGE FOLKSONOMY improving the effectiveness of NDVC detection 1.2 1 Input shot Si 0.8 NDCR Visual Image folksonomy F 0.6 Extraction of low-level visual features Descending order of popularity features User 0.4 Ascending order of popularity User Content-based image retrieval User-contributed images User-supplied tags 0.2 Images User-contributed images k nearest visual neighbors of Si & tags User-supplied tags 0 120 310 10 20 30 40 50 60 70 80 90 100 110 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 320 330 340 350 360 370 : night, sky, stars, mountains, milkyway, aquila, User-contributed images sagittarius, scorpius, ... User-contributed images Number of semantic concepts used User-supplied tags User-supplied tags : milkyway, sky, space, astrophotography, Fig. 2. Influence of semantic concept popularity on NDVC detection. night, telescope, jupiter, clouds, ... User User 2.2. Influence of different types of video content ... - To facilitate effective NDVC detection, video signatures need to be : milky way, galaxy, stars, sky robust against the use of different types of video content - category 1 (documentaries), category 2 (news), category 3 (drama and movies), category 4 (miscellaneous) Fig. 1. Retrieval of the k nearest visual neighbor images and their associated tags from an image folksonomy F for a video shot Si. - The effectiveness of the proposed NDVC detection technique is stable and high for all types of video content investigated - Metric for measuring the relevance of a tag t w.r.t. the shot Si: c : the frequency of t in the set of k neighbors c Lt R (t ) = - , Lt : the number of images labeled with t in F K F F : the number of images in F - Layout of the semantic feature signature Ai of a shot Si: [ ] Ai = ti , j , wi , j , j = 1,..., Ai , wi , j : a weight value for tag ti,j - Computation of the weight value for tag ti,j : R(ti , j ) wi , j = Ai Fig. 3. Effectiveness of NDVC detection for different types of video content. ∑ R(ti, k ) Key frame Model-based approach Model-free approach Key frame Model-based approach Model-free approach k =1 Cloud Stars Sky Night Water Geotagged N/A N/A III. SEMANTIC DISTANCE MEASUREMENT USING THE Moonlight Rainbow Constellation Sky SIGNATURE QUADRATIC FORM DISTANCE (SQFD) … … - Adaptive semantic distance measurement between shots Sq and Sr: She Puppy Dog r T Blue w |- w G w |- w q r q r q r q Dshot (S , S ) = SQFD(A , A ) = Civilian Person Grass , Group Clouds Zoo N/A Summer Safari … … q q q r r r w w ,...,w 1 Aq w w ,...,w 1 Fig. 4. Example key frames with detected semantic concepts Ar (underlined semantic concepts are considered to be correct). V. CONCLUSIONS - The elements of the ground similarity matrix G: -This paper discussed a novel technique for NDVC detection - takes advantage of the collective knowledge in an image folksonomy It i tj I ti ∩ t j : the set of images annotated with both tag ti and tj - allows using an unrestricted and dynamic concept vocabulary gij , - takes advantage of the flexible SQFD metric It I ti : the set of images annotated with tag ti - allows taking into account that the nature, the relevance, and the i number of semantic concepts may strongly vary from shot to shot IEEE International Conference on Multimedia and Expo (ICME), July 2011, Barcelona (Spain)