Video 32: Edge Histogram Descriptor (EHD) of Mpeg-7 for Image Retrieval

Hello Viewers, in this video, Edge Histogram Descriptor (EHD) of MPEG-7 family is explained. It is shown that EHD is one of the effective visual descriptor which focuses on the spatial edge distribution in an image. EHD mainly captures five types of edge orientations such as Vertical, Horizontal, Diagonal 45, Diagonal 135 and isotropic.

This EHD is compact in size and is just length of 80 points. If global bin is also attached then it becomes of size 85. Due to small size, it is suitable for fast image search. In this video, Theory of EHD, its implementation in MATLAB and its effectiveness in image retrieval is shown.

This video has following contents:

  • What is Content Based Image Retrieval (CBIR)? 
  • MPEG-7 Visual Descriptors.
  • Edge Histogram Descriptor (EHD).
  • Steps to compute EHD.
  • MATLAB Code for implementing EHD.
  • Applying EHD on images and interpreting the outcome.
  • Using EHD for image retrieval.

Link of previous video:

Color Layout Descriptor (CLD) of Mpeg-7 for Image Retrieval: Click Here

Download Resources: 

1. All Images: Download

2. Wang Database: Download