Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-12-13 Cooperative journals: 《计算机应用研究》
Abstract: Aiming at the high cost of obtaining the training data set, proposing a new weak supervision method for image saliency detection. Only using the picture-level label when training the network model. Dividing the method into two stages. In the first stage, training the classification model according to the picture-level label to obtain the foreground inference graph. In the second stage, processing the original image by super-pixel block and merged with the foreground inference graph obtained in phase one, thus refining significant object boundaries. The algorithm uses existing large training sets and image-level tags, eliminating the use of pixel-level tags, which reduces the amount of annotation work. The experimental results on the four common benchmark datasets show that the performance is significantly better than the unsupervised model, and it has certain advantages compared with the full-supervised model.