Video scene categorization by 3D hierarchical histogram matching

Paritosh Gupta, Sai Sankalp Arrabolu, Matthew Brown, Silvio Savarese

Research output: Contribution to conferencePaperpeer-review

11 Citations (SciVal)
181 Downloads (Pure)


In this paper we present a new method for categorizing video sequences capturing different scene classes. This can be seen as a generalization of previous work on scene classification from single images. A scene is represented by a collection of 3D points with an appearance based codeword attached to each point. The cloud of points is recovered by using a robust SFM algorithm applied on the video sequence. A hierarchical structure of histograms located at different locations and at different scales is used to capture the typical spatial distribution of 3D points and codewords in the working volume. The scene is classified by SVM equipped with a histogram matching kernel, similar to [21, 10, 16]. Results on a challenging dataset of 5 scene categories show competitive classification accuracy and superior performance with respect to a state-of-the-art 2D pyramid matching methods [16] applied to individual image frames.
Original languageEnglish
Number of pages8
Publication statusPublished - Sept 2009
EventICCV 2009: IEEE 12th International Conference on Computer Vision - Kyoto
Duration: 29 Sept 20092 Oct 2009


ConferenceICCV 2009: IEEE 12th International Conference on Computer Vision


Dive into the research topics of 'Video scene categorization by 3D hierarchical histogram matching'. Together they form a unique fingerprint.

Cite this