Domain Invariant Siamese Attention Mask for Small Object Change Detection via Everyday Indoor Robot Navigation
This addresses the problem of small object change detection for robotics applications, offering an incremental improvement by adapting existing attention mechanisms to reduce retraining costs.
The paper tackles the challenge of detecting small, visually non-distinctive changes in images during indoor robot navigation by proposing a novel self-attention technique with unsupervised on-the-fly domain adaptation, which significantly boosts the state-of-the-art image change detection model in experiments.
The problem of image change detection via everyday indoor robot navigation is explored from a novel perspective of the self-attention technique. Detecting semantically non-distinctive and visually small changes remains a key challenge in the robotics community. Intuitively, these small non-distinctive changes may be better handled by the recent paradigm of the attention mechanism, which is the basic idea of this work. However, existing self-attention models require significant retraining cost per domain, so it is not directly applicable to robotics applications. We propose a new self-attention technique with an ability of unsupervised on-the-fly domain adaptation, which introduces an attention mask into the intermediate layer of an image change detection model, without modifying the input and output layers of the model. Experiments, in which an indoor robot aims to detect visually small changes in everyday navigation, demonstrate that our attention technique significantly boosts the state-of-the-art image change detection model.