CNN-RNN: A Unified Framework for Multi-label Image Classification

TLDR

Deep CNNs excel at single‑label image classification, but real‑world images often contain multiple labels, and existing multi‑label methods train independent classifiers that ignore label dependencies. This work proposes using recurrent neural networks to capture label dependencies in multi‑label image classification. The CNN‑RNN framework jointly learns an image‑label embedding that models semantic label dependencies and image‑label relevance, and is trained end‑to‑end from scratch. Experiments on public benchmarks show the architecture outperforms state‑of‑the‑art multi‑label models.

Abstract

While deep convolutional neural networks (CNNs) have shown a great success in single-label image classification, it is important to note that real world images generally contain multiple labels, which could correspond to different objects, scenes, actions and attributes in an image. Traditional approaches to multi-label image classification learn independent classifiers for each category and employ ranking or thresholding on the classification results. These techniques, although working well, fail to explicitly exploit the label dependencies in an image. In this paper, we utilize recurrent neural networks (RNNs) to address this problem. Combined with CNNs, the proposed CNN-RNN framework learns a joint image-label embedding to characterize the semantic label dependency as well as the image-label relevance, and it can be trained end-to-end from scratch to integrate both information in a unified framework. Experimental results on public benchmark datasets demonstrate that the proposed architecture achieves better performance than the state-of-the-art multi-label classification models.