230 likes | 353 Views
Giving Meanings to WWW Images. Heng Tao Shen Beng Chin Ooi Kian Lee Tan. Outline. Image Representation Model Semantic Measure Model Relevance Feedback Experiments. Background. Image: indispensable component in WWW 1 image = 1000 words WWW: rich resource of images Some 100 billions?
E N D
Giving Meanings to WWW Images Heng Tao Shen Beng Chin Ooi Kian Lee Tan ACM MM 2000, LA, USA
Outline • Image Representation Model • Semantic Measure Model • Relevance Feedback • Experiments
Background • Image: indispensable component in WWW • 1 image = 1000 words • WWW: rich resource of images • Some 100 billions? • Tradition: poor performance • Keywords • Content_based: no enough semantic • Like object, event, and relationship Not effective for images from WWW
Cont • Semantics of embedded images in HTML • Image Title, ALT, Page Title, Image Caption -> ChainNet model • Similarity between query and image • List space model • Relevance feedback: • Improve precision further
Weight ChainNet model • Lexical chain(LC) • A sentence that carries certain semantics by its words • 6 types of LC • TLC: Title Lexical Chain • PLC: Page Lexical Chain • ALC: Alt Lexical Chain • SLC: Sentence Lexical Chain • RSLC: Reconstructed Sentence Lexical Chain • CLC: Caption Lexical Chain
Page Title Title ALT Caption 4 1 7 2 SLC: 1->2->3->4->5 RSLC: 1->2->8->9 CLC: 1->2->…->14 8 3 9 14 4 5
Semantic measure model • Computing similarity between two LCs • List space model Where ei and ej are matched terms in list 1 and list 2 respectively.
Semantic measure model • Match scale: closeness in view of match order Here v1 and v2 represent the children of first and second original lists respectively. Where v2j is the matched word in v2 for v1i in v1 Inspired from the angle between two vectors
Semantic measure model LC Match Level(LC1, LC2): the number of distinct matched words by two LCs • Match level threshold: The minimum match level for LC to keep its original semantic • LC Semantic similarity: similarity(list1, list2) in its LC Match Level
Semantic measure model Image Match Level(image, query) = MAX ( TLC.weight * LCMatchLevel( TLC, QLC), ALC.weight * LCMatchLevel( ALC, QLC), PLC.weight * LCMatchLevel( PLC, QLC), SLC.weight * LCMatchLevel( SLC, QLC), RSLC.weight *LCMatchLevel( RSLC, QLC), CLC.weight * LCMatchLevel( CLC, QLC) )
Relevance Feedback • Semantic Accumulation • Choose one best image as feedback • Accumulate the previous feedback images’ semantics to construct a new QLC • Results are more close to the specific image selected • More noise
New query QLC Image Title Page Title Image Caption Image ALT Last feedback image Semantic accumulation Weight F/Q ChainNet
Semantic Integration and Differentiation • Semantic Integration and Differentiation • Choose several Good and Bad images as feedback • Integrate Good semantics to construct new query • Differentiate irrelevant images by Bad images • Results are more diverse and less noise
Most related LC New query QLC LC1 LC2 LC3 LCi Good feedback images Image 1 Image 2 Image 3 Image i Semantic integration and differentiation Similar weight F/Q ChainNet
Experiments • Set up • Web crawler to collect images • 5232 images from over 2000 URLs • 12 general queries
Tune the match level MatchLevel Threshold= coef * query.length()+ constant
Impact of match scale • explore the importance of match order
Feedback Mechanisms One-step feedback of Accu and I&D for Q1.
Conclusion • Inner semantic structure of surrounding text is explored well for good precision achievement • ChainNet model and list space model work well • RF techniques help to return more accurate results
Future work • Explore LC meanings by AI technique • Extract semantics from visual content, then integrate with our system to construct a more advanced semantic retrieval system • Object-oriented detection
DEMO ON THURSDAY SEE YOU THEN… http://efoto.geofoto.com