 |
LOCAL SEARCH
Viewgraphs with graphics can be found at
http://www-db.stanford.edu/pub/gio/personal/WIPE-NRC.ppt
Presentation below was delivered by Michel Bilello, Stanford University
Advanced Techniques for Automatic Web Filtering
James Z. Wang
PNC Tech. Career Dev. Professor
Penn State University
Joint Work: Jia Li, Assist. Prof., Penn State Statistics
Gio Wiederhold, Prof., Stanford Computer Science
Outline
- The problem
- Related approaches
- Filtering based on image content
- Goals and methods
- The WIPE system
- Experimental results
- Web site classification by image content
- Conclusions and future work
The Size and Content of the Web
02/99: 16 million total Web servers
- Estimated total number of pages on the Web: 800 million
- 15 Terabytes of text - comparable to text of Library of Congress)
- Year 2001: 3 to 5 billion pages
Lawrence, Giles, Nature, 1999.
Outline
- The problem
- Related approaches
- Filtering based on image content
- Goals and methods
- The WIPE system
- Experimental results
- Web site classification by image content
- Conclusions and future work
Pornography-free Web sites
- E.g. Yahoo!Kids, disney.com
- Useful in protecting those children too young to know how to use the Web browser
- It is difficult to control access to other sites
Text-based Filtering
- E.g. NetNanny, Cyber Patrol, CyberSitter
- Methods:
- Store more than 10,000 IPs
- Blocking based on keywords
- Block all image access
- Problems:
- Internet is dynamic
- Keywords are not enough (e.g. text incorporated in images)
- Images are needed for all net users
Classification of Web Community
- Flake, Lawrence, Giles, ACM KDD, 2000
- Graph clustering based on max flow – min cut analysis of the Web connectedness
Outline
- The problem
- Related approaches
- Filtering based on image content
- Goals and methods
- The WIPE system
- Experimental results
- Web site classification by image content
- Conclusions and future work
Goals and Methods
- The problem comes from images, we deal with images
- Goals: use machine learning and image retrieval to classify Web images and Web sites
- Requirements: high accuracy and high speed
- Challenges: non-uniform image background, textual noise in foreground, wide range of image quality, wide range of camera positions, wide range of composition…
The WIPE System
- Inspired by the UC Berkeley’s FNP System
- Detailed analysis of images
- Skin filter and human figure grouper
- Speed: 6 mins CPU time per image
- Accuracy: 52% sensitivity and 96% specificity
- Stanford WIPE System
- Wavelet-based feature extraction + image classification + integrated region matching + machine leaning
- Speed: < 1 second CPU time per image
- Accuracy: 96% sensitivity and 91% specificity
System Flow
Wavelet Principle
Type Classification
Graphs:
Manually-generated images with smooth tones.
Type Classification
Photographs:
Images with continuous tones.
Photo Classification
Experimental Results
- Tested on a set of over 10,000 photographic images
- Speed: Less than one second of response time on a Pentium III PC
- Accuracy
Comment on Accuracy
- The algorithm can be adjusted to trade off specificity for higher sensitivity
- In a real-world filtering application system, both the sensitivity and the specificity are expected to be higher
- Icons and graphs can be classified with almost 100% accuracy higher specificity
- Combine text and image classification higher sensitivity and higher speed
False Classifications
Benign Images
False Classifications
Objectionable Images
Web site Classification
by Image Content
- An objectionable site will have many such images
- For a given objectionable Web site, we denote p as the chance of an image on the Web site to be an objectionable image
- p is the percentage of objectionable images over all images provided by the site
- We assume some distributions of p over all Web sites (e.g., Gaussian, shifted Gaussian)
- Classification levels could be provided as a service to filtering software producers
Flow in Web site classification
Web site Classification
- Based on statistical analysis (see paper), we know we can expect higher than 97% accuracy on Web site classification if
- We download 20-35 images for each site
- We classify a Web site as objectionable if 20-25% of downloaded images are objectionable
- Using text and IP addresses as criteria, the accuracy can be further improved
- skip IPs for museums, dog-shows, beach towns, sport events
Outline
- The problem
- Related approaches
- Filtering based on image content
- Goals and methods
- The WIPE system
- Experimental results
- Web site classification by image content
- Conclusions and future work
Conclusions and Future Work
- Perfect filtering is never possible
- Effective filtering based on image content is feasible with the current technology
- Systems that combine content-based filtering with text-based criteria will have good accuracy and acceptable speed
- Objectionable Web sites are automatically identifiable, a service for the community?
- The technology can still be improved through further research.
References
- http://www-db.stanford.edu/IMAGE (papers)
- http://wang.ist.psu.edu/cgi-bin/zwang/wipe2_show.cgi(demo)
- http://www-db.stanford.edu/pub/gio/inprogress.html - COPA(testimony)
- jwang@ist.psu.edu(James Wang)
- gio@cs.stanford.edu (Gio Wiederhold)
- michel@db.stanford.edu(Michel Bilello)
