Browse Prior Art Database

Method and System of Detecting Fake Advertisements through Image Recognition

IP.com Disclosure Number: IPCOM000241871D
Publication Date: 2015-Jun-05
Document File: 2 page(s) / 115K

Publishing Venue

The IP.com Prior Art Database

Related People

Angus Xianen Qiu: INVENTOR [+3]

Abstract

A method and system for detecting fake advertisements (ads) or creative mimics through image recognition is disclosed. The method and system utilizes a training model to classify fake ads that might exist in webpages.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 63% of the total text.

Method and System of Detecting Fake Advertisements through Image Recognition

Abstract

A method and system for detecting fake advertisements (ads) or creative mimics through image recognition is disclosed.  The method and system utilizes a training model to classify fake ads that might exist in webpages.

Description

A method and system is disclosed for detecting fake advertisements (ads) or creative mimics through image recognition.  The method and system utilizes a training model to identify fake ads that might exist in webpages.

In an implementation, the method and system involves a training phase and a prediction phase to identify fake ads as illustrated in fig. 1. 

Figure 1

Initially in training phase, image snapshots of one or more ads are created.  Subsequently, one or more features such as a sift, a surf, a hog and a cbow are extracted by analyzing the image.  Further, a codebook is generated by utilizing the one or more extracted features of image.  Thereafter, a feature histogram is calculated for each image using the codebook.  Finally, a model is trained using the feature histograms associated with labels to classify images. 

Moving on, during prediction phase, fake ads are identified using a training model. In order to do so, an image snapshot is taken for an ad (potentially a fake ad).  Thereafter one or more features are extracted from the image.  Subsequently, a feature histogram of the image is calculated using a codebook generated in training phase....