Object Recognition using Novel Geometrical Feature Extraction Techniques

In Image Processing, an object is an identifiable portion of a particular image that can be interpreted as a single unit. Humans have the ability to recognize any type of objects whether they are alphabets, digits or any living and non-living things irrespective of their forms. When it comes to a machine, it detects an object by extracting its features. Feature Extraction is the most popular research area in the field of image analysis, and it is the primary requirement for representing an object. By these feature extraction techniques, the objects will be represented as a group of features in the form of feature vectors and then they are used for the recognition of objects and for classifying them. In this paper, we have proposed geometrical features from the set of training images using triangular area and perimeter. These features of the training images are stored in the database and used for classifying the test images and Chi-Square statistics is used as classification method