POCA-Lite: A Lightweight Change-Detection Architecture with Geometry-Aware Auxiliary Supervision and Feedback Fusion.
Saved in:
| Title: | POCA-Lite: A Lightweight Change-Detection Architecture with Geometry-Aware Auxiliary Supervision and Feedback Fusion. |
|---|---|
| Authors: | Shi, Yongqi1 (AUTHOR) shiyongqi17@nudt.edu.cn, Yang, Ruopeng2 (AUTHOR), Huang, Bo1 (AUTHOR), Gu, Zhaoyang1,2 (AUTHOR), Lu, Yiwei2 (AUTHOR), Yin, Changsheng2 (AUTHOR), Wen, Yongqi1 (AUTHOR), Zhong, Yihao1 (AUTHOR) |
| Source: | Remote Sensing. May2026, Vol. 18 Issue 10, p1673. 35p. |
| Subjects: | Change-point problems, Artificial neural networks, Multisensor data fusion, Remote sensing |
| Abstract: | Highlights: What are the main findings? POCA-lite, a 1.33 M-parameter encoder–decoder with an inference-coupled geometry branch, matches SNUNet in mean F1 on LEVIR-CD while using 47% fewer parameters and 53% fewer FLOPs. Decomposition ablations disentangle two complementary gain sources: geometric supervision alone recovers 85% of the total improvement, while the feedback fusion pathway alone recovers 92%; their combination achieves the full result. Boundary F1 improves by 9.22 percentage points over the no-geometry baseline, and cross-architecture transfer to SNUNet yields +1.06 pp F1. What are the implications of the main findings? Inference-coupled geometric supervision is a promising strategy for lightweight, boundary-sensitive building change detection on domains with well-separated morphology. Cross-dataset evaluation on WHU-CD reveals that the geometric assumptions degrade on dense, irregular building layouts, establishing clear scope boundaries and guiding practitioners on when to apply or avoid this approach. Building change detection from bi-temporal remote-sensing imagery underpins urban planning, infrastructure monitoring, and disaster assessment. Existing deep-learning methods achieve high accuracy but rely on large parameter counts, while pixel-level supervision provides limited boundary guidance. We propose POCA-lite, a lightweight encoder–decoder with an inference-coupled geometry branch: three geometric prediction heads—distance transform, boundary, and center heatmap—whose outputs are fused back into the decoder via a feedback pathway active at both training and inference. On the LEVIR-CD benchmark under a unified retraining protocol, multi-seed evaluation shows that POCA-lite matches SNUNet in mean F1 while using 47% fewer parameters and 53% fewer FLOPs. Boundary F1 improves by 9.22 pp over the no-geometry baseline. Decomposition ablations reveal two complementary improvement sources: geometric supervision alone recovers 85% of the total gain, while the feedback fusion pathway recovers 92%; their combination achieves the full result. Geometry-aware targets outperform a generic multitask control. Cross-architecture transfer to SNUNet yields +1.06 pp F1. However, cross-dataset evaluation on WHU-CD shows that the method underperforms SNUNet on dense urban morphology, and zero-shot cross-dataset transfer is not established. These results indicate that inference-coupled geometric supervision is effective for lightweight, boundary-sensitive change detection on domains with well-separated building morphology, but its applicability is scope-bounded. [ABSTRACT FROM AUTHOR] |
| Copyright of Remote Sensing is the property of MDPI and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | Highlights: What are the main findings? POCA-lite, a 1.33 M-parameter encoder–decoder with an inference-coupled geometry branch, matches SNUNet in mean F1 on LEVIR-CD while using 47% fewer parameters and 53% fewer FLOPs. Decomposition ablations disentangle two complementary gain sources: geometric supervision alone recovers 85% of the total improvement, while the feedback fusion pathway alone recovers 92%; their combination achieves the full result. Boundary F1 improves by 9.22 percentage points over the no-geometry baseline, and cross-architecture transfer to SNUNet yields +1.06 pp F1. What are the implications of the main findings? Inference-coupled geometric supervision is a promising strategy for lightweight, boundary-sensitive building change detection on domains with well-separated morphology. Cross-dataset evaluation on WHU-CD reveals that the geometric assumptions degrade on dense, irregular building layouts, establishing clear scope boundaries and guiding practitioners on when to apply or avoid this approach. Building change detection from bi-temporal remote-sensing imagery underpins urban planning, infrastructure monitoring, and disaster assessment. Existing deep-learning methods achieve high accuracy but rely on large parameter counts, while pixel-level supervision provides limited boundary guidance. We propose POCA-lite, a lightweight encoder–decoder with an inference-coupled geometry branch: three geometric prediction heads—distance transform, boundary, and center heatmap—whose outputs are fused back into the decoder via a feedback pathway active at both training and inference. On the LEVIR-CD benchmark under a unified retraining protocol, multi-seed evaluation shows that POCA-lite matches SNUNet in mean F1 while using 47% fewer parameters and 53% fewer FLOPs. Boundary F1 improves by 9.22 pp over the no-geometry baseline. Decomposition ablations reveal two complementary improvement sources: geometric supervision alone recovers 85% of the total gain, while the feedback fusion pathway recovers 92%; their combination achieves the full result. Geometry-aware targets outperform a generic multitask control. Cross-architecture transfer to SNUNet yields +1.06 pp F1. However, cross-dataset evaluation on WHU-CD shows that the method underperforms SNUNet on dense urban morphology, and zero-shot cross-dataset transfer is not established. These results indicate that inference-coupled geometric supervision is effective for lightweight, boundary-sensitive change detection on domains with well-separated building morphology, but its applicability is scope-bounded. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 20724292 |
| DOI: | 10.3390/rs18101673 |
