ViewDelta: Scaling Scene Change Detection through Text-Conditioning
This addresses the problem of generalizing SCD across diverse domains for applications like urban monitoring or autonomous systems, though it is incremental by building on existing methods with a novel conditioning approach.
The paper tackles the ambiguity in Scene Change Detection (SCD) by introducing ViewDelta, a text-conditioned framework that uses natural language prompts to define relevant changes, and it achieves performance competitive with or superior to dataset-specific models across multiple datasets.
We introduce a generalized framework for Scene Change Detection (SCD) that addresses the core ambiguity of distinguishing "relevant" from "nuisance" changes, enabling effective joint training of a single model across diverse domains and applications. Existing methods struggle to generalize due to differences in dataset labeling, where changes such as vegetation growth or lane marking alterations may be labeled as relevant in one dataset and irrelevant in another. To resolve this ambiguity, we propose ViewDelta, a text conditioned change detection framework that uses natural language prompts to define relevant changes precisely, such as a single attribute, a specific set of classes, or all observable differences. To facilitate training in this paradigm, we release the Conditional Change Segmentation dataset (CSeg), the first large-scale synthetic dataset for text conditioned SCD, consisting of over 500,000 image pairs with more than 300,000 unique textual prompts describing relevant changes. Experiments demonstrate that a single ViewDelta model trained jointly on CSeg, SYSU-CD, PSCD, VL-CMU-CD, and their unaligned variants achieves performance competitive with or superior to dataset specific models, highlighting text conditioning as a powerful approach for generalizable SCD. Our code and dataset are available at https://joshuakgao.github.io/viewdelta/.