ASK: Adaptively Selecting Key Local Features for RGB-D Scene Recognition

TL;DR

This paper introduces an adaptive local feature selection framework for RGB-D scene recognition, effectively capturing spatial variability and improving classification accuracy by selecting key features from multi-modal data.

Contribution

It proposes a differentiable local feature selection module that adaptively chooses discriminative local features using mutual information maximization, enhancing RGB-D scene classification.

Findings

Achieves state-of-the-art performance on public datasets.

Effectively captures spatial variability with adaptive feature selection.

Utilizes RGB-D modality correlation for improved feature discrimination.

Abstract

Indoor scene images usually contain scattered objects and various scene layouts, which make RGB-D scene classification a challenging task. Existing methods still have limitations for classifying scene images with great spatial variability. Thus, how to extract local patch-level features effectively using only image labels is still an open problem for RGB-D scene recognition. In this paper, we propose an efficient framework for RGB-D scene recognition, which adaptively selects important local features to capture the great spatial variability of scene images. Specifically, we design a differentiable local feature selection (DLFS) module, which can extract the appropriate number of key local scenerelated features. Discriminative local theme-level and object-level representations can be selected with the DLFS module from the spatially-correlated multi-modal RGB-D features. We take advantage of the correlation between RGB and depth modalities to provide more cues for selecting local features. To ensure that discriminative local features are selected, the variational mutual information maximization loss is proposed. Additionally, the DLFS module can be easily extended to select local features of different scales. By concatenating the local-orderless and global structured multi-modal features, the proposed framework can achieve state-of-the-art performance on public RGB-D scene recognition datasets.