Object detection using shape features
Shalu Gupta, Yumnam Jayanta Singh · 2014
This study presents a method for detecting an object by using several features. Features are extracted based on the statistical distribution of points on the bitmap image of the shape. It detect the shape of an image by using geometrical features and then select the clustered of image that matched that shape from a large database. In this approach the image of object are divided into zones and find out some feature of each zone. The extracted features are used for classification based on zones. To make a decision for classification the feature of zone are extracted as per the boundary structure segment-the sum of ON or foreground pixels in each boundary segment, an angle between the starting and ending points, a distance between the starting and ending points, centroid of each zone, an angle between the starting and centroid points, a distance between the starting and centroid points, an angle between the centroid and ending points, a distance between the centroid and ending points etc.