Describing and Localizing Multiple Changes with Transformers
This work addresses the limitation of existing change captioning methods that focus on single changes, enhancing adaptability for applications in complex visual scenarios, though it is incremental as it builds on prior single-change captioning approaches.
The paper tackles the problem of detecting and describing multiple changes in image pairs, which is essential for complex scenarios, by proposing a new dataset, benchmarking existing methods, and introducing Multi-Change Captioning transformers (MCCFormers) that achieve the highest scores on four evaluation metrics and outperform previous state-of-the-art methods by large margins (e.g., +6.1 on BLEU-4 and +9.7 on CIDEr).
Change captioning tasks aim to detect changes in image pairs observed before and after a scene change and generate a natural language description of the changes. Existing change captioning studies have mainly focused on a single change.However, detecting and describing multiple changed parts in image pairs is essential for enhancing adaptability to complex scenarios. We solve the above issues from three aspects: (i) We propose a simulation-based multi-change captioning dataset; (ii) We benchmark existing state-of-the-art methods of single change captioning on multi-change captioning; (iii) We further propose Multi-Change Captioning transformers (MCCFormers) that identify change regions by densely correlating different regions in image pairs and dynamically determines the related change regions with words in sentences. The proposed method obtained the highest scores on four conventional change captioning evaluation metrics for multi-change captioning. Additionally, our proposed method can separate attention maps for each change and performs well with respect to change localization. Moreover, the proposed framework outperformed the previous state-of-the-art methods on an existing change captioning benchmark, CLEVR-Change, by a large margin (+6.1 on BLEU-4 and +9.7 on CIDEr scores), indicating its general ability in change captioning tasks.