Color Objects Recognition System based on Artificial Neural Network with Zernike, Hu & Geodesic Descriptors

In this paper, we propose a hybrid approach based on neural networks and the combination of the classic Hu & Zernike moments joined with Geodesic descriptors. To be able to keep the maximum amount of information that are given by the color of the image, we have calculated Zernike & Hu for each color level. On the other side, geodesic descriptors are applied directly to binary images, and so we can have more information about the general shape of the object. The extracted vectors are put together to form a unique input data to the Neural network. The experimental results showed that the recognition rate of the ANN shape recognition based on the combination of Hu, Zernike & Geodesic descriptors results are noticeably improved. It is also important to note the robustness of the proposed system against the existence of noise, the luminance change, and geometric distortion.