Workflow system for detection and classification of images suspected as pornographic

• US 6,904,168 B1
• Filed: 10/22/2001
• Issued: 06/07/2005
• Est. Priority Date: 03/29/2001
• Status: Expired due to Term

• Current Assignee: Cedar Lane Technologies Inc.
• Original Assignee: FotoNation Holdings, LLC (Adeia Inc.)
• Inventors: Raducan, Ilariu, Steinberg, Eran, Bigioi, Petronel
• Primary Examiner(s): Johns, Andrew W.
• Assistant Examiner(s): Alavi, Amir
• Application Number:
  US10/040,901
• Time in Patent Office: 1,324 Days
• Field of Search: 382/100, 382/162, 382/165, 382/181, 382/189, 382/203, 382/224, 382/225, 382/266, 382/305, 709217-219, 709224-225, 709/229, 709/232
• US Class Current: 382/165
• CPC Class Codes: G06T 7/0002 Inspection of images, e.g. flaw detection
• CPC Class Codes: G06T 7/70 Determining position or orientation of objects or cameras camera calibration G06T7/80
• CPC Class Codes: G06T 7/90 Determination of colour characteristics
• CPC Class Codes: G06V 40/161 Detection; Localisation; Normalisation
• CPC Class Codes: G06V 40/171 Local features and components; Facial parts eye characteristics G06V40/18; Occluding parts, e.g. glasses; Geometrical relationships
• CPC Class Codes: G06V 40/193 Preprocessing; Feature extraction
• CPC Class Codes: H04N 1/624 Red-eye correction control of camera based on recognised facial parts H04N23/611

Abstract

An enterprise based workflow server based system to detect and classify images as potentially of pornographic nature. The system is based on two independent stages. The first stage is an automatic stage which eliminates most images that are positively non pornographic, including an image analysis engine with a plurality of sub engines, each dedicated to different attributes of an image. The second stage describes an additional embodiment of a manual review stage. An alternative embodiment describes a similar server solution to detect images suspected as copyrighted material. The images can be individually classified, or analyzed as a cluster or a collection of images such as an album or motion-picture segment, sampled into key still frames. The system optimizes time saving by elimination of images, and assures accuracy by generating a workflow reviewing solution which can involve a human decision-making stage.

Claims

1. A server based system for classification of images suspected as pornographic comprising:

(a) a first defined interface between a client application and said server for transferring images and related meta data;

(b) a database for storing and queuing the images and meta data associated with said image;

(c) independent engine apparatus for automatically analyzing the images using a plurality of independent review engines for a plurality of criteria for analyzing results from said plurality of independent review engines of said images into a single parameter representing a likelihood value that an image is of pornographic nature; and

(d) a second defined interface to convey a result of said analyzing to said client application, said result including a list of most suspected images.

2. A server based system for classification of images suspected as pornographic as recited in claim 1 wherein said automatic analyzing eliminates a majority of said images and is performed using color analysis on skin tone.

3. A server based system for classification of images suspected as pornographic as recited in claim 1 wherein said automatic analyzing eliminates a majority of said images and is performed using a shape analysis.

4. A server based system for classification of images suspected as pornographic as recited in claim 1 wherein said automatic analyzing eliminates a majority of said images and is performed by means of a curvature analysis.

5. A server based system for classification of images suspected as pornographic as recited in claim 1 wherein said automatic analyzing eliminates a majority of said images and is performed by texture analysis.

6. A server based system for classification of images suspected as pornographic as recited in claim 1 wherein said automatic analyzing eliminates a majority of said images and is performed by means of body pose and posture analysis.

7. A server based system for classification of images suspected as pornographic as recited in claim 1 wherein said automatic analyzing eliminates a majority of said images and is performed by means of a combination of color analysis on skin tone, shape analysis, curvature analysis, texture analysis, body pose and posture analysis.

8. A server based system for classification of images suspected as pornographic as recited in claim 1 wherein said images are clustered as a collection and said analyzing is performed on said cluster.

9. A server-based system for classification of images suspected as pornographic as recited in claim 8 wherein said analyzing of said cluster includes determining a collective likelihood of the cluster based on combined statistical estimation of individual images in the cluster.

10. A server-based system for classification of images suspected as pornographic as recited in claim 8 wherein said cluster is a continuous collection of individual frames, which creates a single motion-picture clip.

11. A server based system for classification of images suspected as pornographic as recited in claim 10 wherein a selection of said individual frames from said cluster for said classification is determined by automatic detection of a change of scene in said motion picture clip.

12. A server based system for classification of images suspected as pornographic as recited in claim 10 wherein a selection of said individual frames from said cluster for said classification is determined by a predefined interval of said individual frames.

13. A server-based system for classification of images suspected as pornographic as recited in claim 8 wherein said cluster is a continuous collection of motion-picture clip.

14. A server based system for classification of images suspected as pornographic as recited in claim 1 further comprising:

(a) a first apparatus for performing an additional automated step to eliminate a majority of said images that are positively non-pornographic based on said likelihood value;

and

(b) a second apparatus for performing a manual review performed by human reviewers to determine whether said most suspected images are pornographic.

15. A server based system for classification of images suspected as pornographic as recited in claim 14 further comprising an apparatus for creating a queue of said most suspected images for said manual review based on a priority criteria.

16. A server based system for classification of images suspected as pornographic as recited in claim 15 wherein said priority criteria is a number of page views requested for the image.

17. A serve based system for classification of images suspected as pornographic as recited in claim 15 wherein said priority criteria is a statistical likelihood of a said image being pornographic.

18. A server based system for classification of images suspected as pornographic as recited in claim 14 wherein said reviewers assign a rating to each image.

19. A server based system for classification of images suspected as pornographic as recited in claim 14 wherein said reviewers assign a rating to a cluster of images.

20. A server based system for classification of images suspected as pornographic as recited in claim 14 wherein said reviewers are supervised by a supervisor.

21. A server-based system for classification of images suspected as pornographic as recited in claim 20 wherein said reviewers can escalate images that they are uncertain of to said supervisor for determining if said images that said reviewers are uncertain of are pornographic.

22. A server-based system for classification of images suspected as pornographic as recited in claim 14 further comprising apparatus for creating a queue of a collection of images based on a priority criteria.

23. A server based system for classification of images suspected as pornographic as recited in claim 22 wherein said priority criteria is the percent of said suspected images in said collection of images.

24. A server based system for classification of images suspected as pornographic as recited in claim 22 wherein said priority criteria is the total number of said suspected images within said collection of images.

25. A server based system for classification of images suspected as pornographic as recited in claim 1 wherein said images are sub sampled to the smallest image size available which maintains sufficient data for said classification thus saving on transmission time memory and bandwidth.

26. A server based system for classification of images suspected as pornographic as recited in claim 1 wherein said client application is run on a remote computer, said images suspected as pornographic obtained from a data storage subsystem of said computer, and said system further comprising:

(a) a first apparatus for performing an additional automated step to eliminate most images that are positively non-pornographic based on said likelihood value thus maintaining only small numbers of suspected images from said analyzing;

(b) a second apparatus for submitting images for a manual review step performed by a human reviewer to determine whether said suspected images from said automated step are pornographic; and

(c) a third apparatus for performing a clean-up step to allow the deletion of the objectionable pornographic images from said remote computer.

27. A server based system for classification of copyrighted images comprising:

(a) a first-defined interface between a client application and a server to transfer images and related meta data;

(b) a database to store and queue the images and said related meta data;

(c) independent engine apparatus to automatically analyze the images using a plurality of criteria related to copyright material;

(d) apparatus for analyzing the individual reviews into a single parameter representing a likelihood value that an image is copyrighted;

and

(e) a second-defined interface for conveying a result to said client application.

28. A server based system for classification of copyrighted images as recited in claim 27 wherein said criteria include watermarking.

29. A server based system for classification of copyrighted images as recited in claim 27 wherein said copyright criteria include an existence of printed half tone patterns.

30. A server based system for classification of copyrighted images as recited in claim 27 wherein said copyright criteria include an existence of textual phrases that depict celebrity images.

31. A server based system for classification of copyrighted images as recited in claim 27 wherein said copyright criteria include an existence of textual phrases that depict current events.

32. A server based system for classification of copyrighted images as recited in claim 27 wherein said copyright criteria include a binary original of a copyright image.

33. A server based system for detection of graphically-offensive material in images comprising:

(a) a first-defined interface between a client application and a server to transfer images and related meta data;

(b) a database to store and queue the images and said related meta data;

(c) independent engine apparatus to automatically analyze the images using a plurality of criteria related to said graphically offensive material;

(d) apparatus for analyzing the individual reviews into a single parameter representing a likelihood value that an image includes said graphically-offensive material;

and

(e) a second-defined interface for conveying a result to said client application.

34. A server based system for detection of graphically-offensive material in images as recited in claim 33 wherein said criteria include the detection of known logos associated with politically adverse organizations.

35. A server based system for detection of graphically-offensive material in images as recited in claim 33 wherein said criteria include the detection of known logos associated with hate crimes.

36. A server based system for detection of graphically-offensive material in images as recited in claim 33 wherein said images include line art representation.

37. A server based system for detection of graphically-offensive material in images as recited in claim 33 wherein said images include vector representation.

38. A server based system for detection of graphically-offensive material in images as recited in claim 33 wherein said images include pixel representation.