Attention to Scale: Scale-Aware Semantic Image Segmentation

TLDR

In semantic image segmentation, incorporating multi‑scale features into fully convolutional networks is essential for state‑of‑the‑art performance, typically achieved by feeding resized inputs and merging features. The study proposes an attention mechanism that learns to softly weight multi‑scale features at each pixel location. The authors adapt a state‑of‑the‑art segmentation model, jointly training it with multi‑scale inputs and the attention mechanism, and validate its performance on PASCAL‑Person‑Part, PASCAL VOC 2012, and MS‑COCO 2014. The attention model outperforms average and max pooling, enables visualization of feature importance across positions and scales, and shows that adding supervision at each scale is essential for achieving excellent performance.

Abstract

Incorporating multi-scale features in fully convolutional neural networks (FCNs) has been a key element to achieving state-of-the-art performance on semantic image segmentation. One common way to extract multi-scale features is to feed multiple resized input images to a shared deep network and then merge the resulting features for pixelwise classification. In this work, we propose an attention mechanism that learns to softly weight the multi-scale features at each pixel location. We adapt a state-of-the-art semantic image segmentation model, which we jointly train with multi-scale input images and the attention model. The proposed attention model not only outperforms averageand max-pooling, but allows us to diagnostically visualize the importance of features at different positions and scales. Moreover, we show that adding extra supervision to the output at each scale is essential to achieving excellent performance when merging multi-scale features. We demonstrate the effectiveness of our model with extensive experiments on three challenging datasets, including PASCAL-Person-Part, PASCAL VOC 2012 and a subset of MS-COCO 2014.