Elham Askari - Qazvin Branch, Islamic Azad University
Amir Masoud Eftekhari Moghadam - Qazvin Branch, Islamic Azad University
Hamidreza Rashidy Kanan

Image segmentation is an important research area in computer vision and many image segmentation methods have been proposed, therefore it is necessary to be able to evaluate theperformance of image segmentation algorithms objectively. In this paper we present a new metric to evaluate the accuracy ofimage segmentation algorithms, based on the most important feature of each segments using neural networks. The neural network after training can assess the similarity or dissimilarity of each pairs of segments, based on the most important feature of two segments that can be distinguished from each other andfinally the segmentation algorithms accuracy have been computed by novel presented metric. Our proposed method donot require a manually-segmented reference image for comparison, therefore can be used for real-time evaluation and is sensitive to over-segmentation. Experimental results were obtained for a selection of images from Berkeley segmentation data set and demonstrated that it’s a proper measure for comparing image segmentation algorithms

objective evaluation; neural network; image segmentation; important feature ; over-segmentation