Image thresholding based on local activity feature matrix

Image thresholding is the simple and widely used image segmentation method in automatic image analysis. In this paper, a novel image thresholding technique based on the local image features is proposed. The core idea is to generate a local activity feature matrix assuming given image a random field. The histogram of the local activity feature matrix, which essentially incorporate the local features of the image, is used for threshold selection process. The performance of the proposed technique is validated by extensive experiments on different real images, where the definite advantages are demonstrated subjectively and objectively. Further, the comparisons with existing techniques confirmed the effectiveness of the proposed technique.