Performance analysis of the Edge Pixel Orientations Histogram

A study on the performance of the Edge Pixel Orientations Histogram is presented in this paper. Edges are detected with the Canny algorithm without and with hysteresis thresholding. The resulting edges are described in No different orientations using the gradient orientation. The two edge images are divided into N × N sub-images. Counting the number of edge pixels with each orientation for each sub-image results in a histogram with 2NoN2 bins. The image descriptor is tested with a subset of the TRECVID 2008 development database k-frames. The resulting Edge Pixel Orientation Histograms will be classified with the K Nearest Neighbour Algorithm and a high level description based on the image semantics is extracted. In particular the concept “images with at least one building” is tested. This descriptor is also used for the JPSearch Alinari image database annotation. Alinari database is composed by a set of 971 high resolution images (3888×2592) shot by professional photographers.

[1]  Miroslaw Bober,et al.  MPEG-7 visual shape descriptors , 2001, IEEE Trans. Circuits Syst. Video Technol..

[2]  Matthijs C. Dorst.  Distinctive Image Features from Scale-Invariant Keypoints , 2011 .

[3]  John F. Canny,et al.  A Computational Approach to Edge Detection , 1986, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[4]  Rainer Lienhart,et al.  The Holy Grail of Multimedia Information Retrieval: So Close or Yet So Far Away? , 2008 .

[5]  Touradj Ebrahimi,et al.  Overview of Jpsearch: A Standard for Image Search and Retrieval , 2007, 2007 International Workshop on Content-Based Multimedia Indexing.

[6]  Tony Lindeberg,et al.  Edge Detection and Ridge Detection with Automatic Scale Selection , 1996, Proceedings CVPR IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[7]  Cordelia Schmid,et al.  A Performance Evaluation of Local Descriptors , 2005, IEEE Trans. Pattern Anal. Mach. Intell..

[8]  Belur V. Dasarathy,et al.  Nearest neighbor (NN) norms: NN pattern classification techniques , 1991 .

[9]  B. S. Manjunath,et al.  Color and texture descriptors , 2001, IEEE Trans. Circuits Syst. Video Technol..

[10]  Antonio Torralba,et al.  Modeling the Shape of the Scene: A Holistic Representation of the Spatial Envelope , 2001, International Journal of Computer Vision.