An Adaptive Color Image Segmentation (A Study and Observations Based on Actual Implementation)

Image segmentation is a primary step in many computer vision tasks. In this paper the method uses watershed algorithm based on gradient magnitude, In which the input RGB color image is transformed into HSI color space and above stated algorithm is applied on the image. Since output of watershed is oversegmented, the results based on actual implementation of adaptive color image segmentation (ACIS) is presented. The implemented system uses a neural network with architecture similar to the multilayer perceptron (MLP) network. The multisigmoid activation function is used for neuron which is the key for segmentation. The neural network is to detect the number of objects automatically from an image.