Fusion based object detection

Fusion based object detection method presented herein combines local and global features for object detection. The cluster based part detection facilitates adequate representation for the individual local parts of an image that carry the significant features of the object. The chain code is used for deriving the local and global information. The genetic algorithm is used to combine all the local and global features. The algorithm is tested on a set of real-world images which contain side views of cars and found to yield very good results with no false alarm within the data base that was used for testing.