Saliency Enhanced Hierarchical Bilinear Pooling for Fine-Grained Classification

Hierarchical bilinear pooling network considers the feature interaction in middle convolutional layers and works well in the classification of fine-grained images.However,it carries out feature interaction on all activated image regions including irrelevant background,which affects the classification performance.To address this problem,a saliency enhanced hierarchical bilinear pooling method is proposed,which combines with the saliency detection network to generate an attention map,and uses the attention map to interact with the feature extraction network to enhance the information of the salient regions.As the result,it can reduce the impact of background and other irrelevant information,and improve the classification performance.The classification accuracy on three commonly used fine-grained image datasets CUB-200-2011,Stanford Cars and FGVC-Aircraft is 86.5%,92.9%and 90.8%,respectively,which is excellent compared with other mainstream methods.