Frequency Based Rearranged Chain Code

This study proposes an algorithm where edge shape information is retrieved using a vote based rearranged chain code. The proposed algorithm rearranges and normalizes the chain code of edges to create a maximum vote based normalization chain code whose correlativity is reinforced. As it obtains cyclical maximum values from the chain code, rearranges them and flips code values according to the frequency, it is resilient against rotation and scaling noise and its retrieval speed is fast. As a result of a test, it was demonstrated that the proposed algorithm can retrieve rotated images better than the existing chain code based algorithm.