An Interactive Image Segmentation Technique Combining Graph Cut And Deep Learning

Image segmentation is one of the most fundamental process in image processing and computer vision communities.The main goal of image segmentation is to divide a digital image into meaningful non-overlapping pixels according to the human perception and to provide further information about the properties of the image data by identifying texture,edges and the regions of similar color of an image.It can benefit various real world application such as content base image retrieval,medical diagnostic,industrial inspection and surveillance.The semi-supervised method,which is also referred as interactive image segmentation involves minimal user interaction,gained more researchers attention in recent years because it can achieve satisfactory results as compared to other state-of-the-art automatic image segmentation methods.Many techniques have demonstrated their potentials for interactive image segmentation.However,most of these state-of-the-art algorithms are unable to produce accurate boundaries without more user interaction,as they are highly sensitive to the seed's quantity and quality.These techniques frequently depend on more user interaction to refine the boundaries.In order to solve these problems,in this work,a research has been made on a robust interactive image segmentation method based on generic multiscale oriented contours via single forward pass Convolutional Neural Networks(CNNs)and graph cut framework.We propose to combine the powerful generic convolutional neural networks with the graph cut framework.We first utilize convolutional neural network to construct the boundary-level information and then combine this boundary-level information with the boundary energy term of graph cut framework to construct a new energy minimization function.The proposed method exhibits better performance in robustness to user interaction,smooth boundaries,accurate segmentation and the ability to handle the changes.To validate the effectiveness of proposed method,we further conducted both qualitative and quantitative experiments on Berkeley segmentation dataset and benchmark and MSRC dataset,showing that our proposed method outperforms the state-of-the-art interactive image segmentation techniques.