Saliency detection method based on active learning

The invention belongs to the technical field of artificial intelligence, and provides a saliency detection method based on active learning. The thought of active learning is applied to the field of saliency detection, a sample which is most beneficial to model training is selected from an unmarked sample set to be added into a training set by considering the uncertainty and diversity of the sample, a final KSR model is obtained through training, and an initial saliency map of a test sample is output by the model. Then, in order to optimize the target boundary of the saliency map, a superpixel-level post-processing method is designed to further improve the performance. According to the method, the labeling cost is reduced, and meanwhile, the redundancy of a training set is reduced, so thatthe experimental effect is greatly improved compared with that of an original KSR model. Meanwhile, a contrast experiment shows that the performance of the method is superior to that of many classicalalgorithms. 本发明属于人工智能技术领域