SCA-CNN: Spatial and Channel-wise Attention in Convolutional Networks for Image Captioning

TL;DR

This paper introduces SCA-CNN, a novel neural network that incorporates spatial and channel-wise attention mechanisms to improve image captioning performance by dynamically focusing on relevant visual features across multiple layers.

Contribution

The paper proposes SCA-CNN, a new CNN architecture that combines spatial and channel-wise attention for enhanced image captioning, addressing limitations of existing spatial-only attention models.

Findings

SCA-CNN outperforms state-of-the-art methods on Flickr8K, Flickr30K, and MSCOCO datasets.

The model effectively encodes where and what to attend in multi-layer features.

Significant performance improvements demonstrate the effectiveness of combined spatial and channel-wise attention.

Abstract

Visual attention has been successfully applied in structural prediction tasks such as visual captioning and question answering. Existing visual attention models are generally spatial, i.e., the attention is modeled as spatial probabilities that re-weight the last conv-layer feature map of a CNN encoding an input image. However, we argue that such spatial attention does not necessarily conform to the attention mechanism --- a dynamic feature extractor that combines contextual fixations over time, as CNN features are naturally spatial, channel-wise and multi-layer. In this paper, we introduce a novel convolutional neural network dubbed SCA-CNN that incorporates Spatial and Channel-wise Attentions in a CNN. In the task of image captioning, SCA-CNN dynamically modulates the sentence generation context in multi-layer feature maps, encoding where (i.e., attentive spatial locations at multiple layers) and what (i.e., attentive channels) the visual attention is. We evaluate the proposed SCA-CNN architecture on three benchmark image captioning datasets: Flickr8K, Flickr30K, and MSCOCO. It is consistently observed that SCA-CNN significantly outperforms state-of-the-art visual attention-based image captioning methods.