Classification of images as advertisement images or non-advertisement images

• US 8,027,940 B2
• Filed: 11/12/2010
• Issued: 09/27/2011
• Est. Priority Date: 06/13/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Li, Mingjing, Li, Dongfang, Wang, Bin, Li, Zhiwei
• Primary Examiner(s): Gaffin; Jeffrey A
• Assistant Examiner(s): CHANG, LI WU
• Application Number:
  US12/945,635
• Publication Number:
  US 20110058734A1
• Time in Patent Office: 319 Days
• Field of Search: 706/12
• US Class Current: 706/12
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0277 Online advertisement

Abstract

An advertisement image classification system trains a binary classifier to classify images as advertisement images or non-advertisement images and then uses the binary classifier to classify images of web pages as advertisement images or non-advertisement images. During a training phase, the classification system generates training data of feature vectors representing the images and labels indicating whether an image is an advertisement image or a non-advertisement image. The classification system trains a binary classifier to classify images using training data. During a classification phase, the classification system inputs a web page with an image and generates a feature vector for the image. The classification system then applies the trained binary classifier to the feature vector to generate a score indicating whether the image is an advertisement image or a non-advertisement image.

Claims

1. A method in a computing device for generating a classifier to identify advertisement images of display pages, the method comprising:

providing training display pages, each training display page having one or more training images referenced by the training display page;

labeling each training image as being an advertisement image or a non-advertisement image;

generating a feature vector for each of the training images, the feature vector including a visual layout feature derived from the display page that references the training mage and a content feature derived from content of the training image; and

training a classifier using the feature vectors and labels of the training images to classify a target image of a target display page as an advertisement image or a non-advertisement image based on a feature vector that includes the visual layout feature derived from the target display page and the content feature derived from content of the target image.

2. The method of claim 1 including classifying the target image by:

generating the feature vector for the target image; and

applying the trained classifier to the generated feature vector of the target image to classify the target image as an advertisement image or a non-advertisement image.

3. The method of claim 1 including learning relevant features for classifying images as advertisement images and non-advertisement images.

4. The method of claim 3 wherein the learning includes applying an adaptive boosting technique to learn weights of weak classifiers, each weak classifier corresponding to a feature, and selecting as relevant features those features whose weights indicate they are effective at distinguishing advertisement images from non-advertisement images.

5. The method of claim 1 wherein features of an image to include in the feature vectors are learned by training a classifier using candidate features and selecting as relevant features those features that are most heavily weighted by the classifier.

6. The method of claim 1 wherein the feature vector of an image further includes a feature derived from a hyperlink.

7. The method of claim 1 wherein the feature vector of an image further includes a feature derived from text of a display page.

8. A computer-readable storage device for classifying a target image of a target display page as an advertisement image or non-advertisement image, by a method comprising:

providing a classifier to classify an image of a display page as an advertisement image or a non-advertisement image, the classifier having been trained using features derived from training images of training display pages and labels of the training images, the label of a training image indicating whether the training image is an advertisement image or a non-advertisement image;

identifying the features of the target image of the target display page; and

applying the provided classifier to the identified features of the target image to classify the target image as an advertisement image or a non-advertisement image.

9. The computer-readable storage device of claim 8 further comprising:

providing training display pages, each training display page having one or more training images referenced by the training display page;

labeling each training image as being an advertisement image or a non-advertisement image;

identifying features for each of the training images; and

training the classifier using the identified features and labels of the training images to classify the target image of the target display page as an advertisement image or a non-advertisement image.

10. The computer-readable storage device of claim 9 including learning relevant features for classifying images as advertisement images and non-advertisement images.

11. The computer-readable storage device of claim 10 wherein the learning includes applying an adaptive boosting technique to learn weights of feature-specific weak classifiers and selecting as relevant features those features of weak classifiers whose weights indicate they are effective at distinguishing advertisement images from non-advertisement images.

12. The computer-readable storage device of claim 8 wherein a feature relates to visual layout of a display page.

13. The computer-readable storage device of claim 8 wherein a feature relates to content of an image.

14. The computer-readable storage device of claim 8 wherein a feature relates to text of a display page.

15. The computer-readable storage device of claim 8 wherein a feature relates to a reference on a display page.

16. The computer-readable storage device of claim 8 wherein a first feature and a second feature are selected from the group consisting of visual layout of a display page feature, content of an image feature, text of a display page feature, and a reference of a display page feature, such that the first feature is different from the second feature.

17. A computing device for identifying features of images of web pages for use in classifying images as advertisement images or non-advertisement images, comprising:

a training data store including feature vectors for images of web pages, the images being labeled as advertisement images or non-advertisement images, the feature vectors including candidate features;

a memory storing computer-executable instructions of;

a component that trains a first classifier using the feature vectors with candidate features and labels of the training data store; and

a component that selects as features for use in classifying images those candidate features whose weights indicate they are effective at distinguishing advertisement images from non-advertisement images; and

a processor for executing the computer-executable instructions stored in the memory.

18. The computing device of claim 17 including a component to train a second classifier using feature vectors of the selected features and labels of the training data store.

19. The computing device of claim 18 including a component that classifies an image as an advertisement image or non-advertisement image by generating a feature vector of the selected features for the image and applying the trained second classifier to the generated feature vector of the image.

20. The computing device of claim 19 wherein a first feature and a second feature are selected from the group consisting of visual layout of a display page feature, content of an image feature, text of a display page feature, and a reference of a display page feature, such that the first feature is different from the second feature.