Mix-Pooling Strategy for Attention Mechanism

Recently many effective self-attention modules are proposed to boot the model performance by exploiting the internal information of convolutional neural networks in computer vision. In general, many previous works ignore considering the design of the pooling strategy of the self-attention mechanism since they adopt the global average pooling for granted, which hinders the further improvement of the performance of the self-attention mechanism. However, we empirically find and verify a phenomenon that the simple linear combination of global max-pooling and global min-pooling can produce pooling strategies that match or exceed the performance of global average pooling. Based on this empirical observation, we propose a simple-yet-effective self-attention module SPENet, which adopts a self-adaptive pooling strategy based on global max-pooling and global min-pooling and a lightweight module for producing the attention map. The effectiveness of SPENet is demonstrated by extensive experiments on widely used benchmark datasets and popular self-attention networks.