Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Cross Branch Feature Fusion Decoder for Consistency Regularization-based Semi-Supervised Change Detection

Yan Xing, Qi'ao Xu, Jingcheng Zeng, Rui Huang, Sihua Gao, Weifeng Xu, Yuxiang Zhang, Wei Fan

TL;DR

A new decoder called Cross Branch Feature Fusion CBFF is introduced, which combines the strengths of both local convolutional branch and global transformer branch, and builds the SSCD model based on a strong-to-weak consistency strategy.

Abstract

Semi-supervised change detection (SSCD) utilizes partially labeled data and a large amount of unlabeled data to detect changes. However, the transformer-based SSCD network does not perform as well as the convolution-based SSCD network due to the lack of labeled data. To overcome this limitation, we introduce a new decoder called Cross Branch Feature Fusion CBFF, which combines the strengths of both local convolutional branch and global transformer branch. The convolutional branch is easy to learn and can produce high-quality features with a small amount of labeled data. The transformer branch, on the other hand, can extract global context features but is hard to learn without a lot of labeled data. Using CBFF, we build our SSCD model based on a strong-to-weak consistency strategy. Through comprehensive experiments on WHU-CD and LEVIR-CD datasets, we have demonstrated the superiority of our method over seven state-of-the-art SSCD methods.

Cross Branch Feature Fusion Decoder for Consistency Regularization-based Semi-Supervised Change Detection

TL;DR

A new decoder called Cross Branch Feature Fusion CBFF is introduced, which combines the strengths of both local convolutional branch and global transformer branch, and builds the SSCD model based on a strong-to-weak consistency strategy.

Abstract

Semi-supervised change detection (SSCD) utilizes partially labeled data and a large amount of unlabeled data to detect changes. However, the transformer-based SSCD network does not perform as well as the convolution-based SSCD network due to the lack of labeled data. To overcome this limitation, we introduce a new decoder called Cross Branch Feature Fusion CBFF, which combines the strengths of both local convolutional branch and global transformer branch. The convolutional branch is easy to learn and can produce high-quality features with a small amount of labeled data. The transformer branch, on the other hand, can extract global context features but is hard to learn without a lot of labeled data. Using CBFF, we build our SSCD model based on a strong-to-weak consistency strategy. Through comprehensive experiments on WHU-CD and LEVIR-CD datasets, we have demonstrated the superiority of our method over seven state-of-the-art SSCD methods.
Paper Structure (10 sections, 12 equations, 4 figures, 2 tables)

This paper contains 10 sections, 12 equations, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Problem formulation
  4. Change Detection Network
  5. Our consistency regularization-based SSCD method
  6. Implementation detail
  7. Experiment
  8. Setup
  9. Results and Discussion
  10. Conclusion

Figures (4)

  • Figure 1: Motivation: Comparison of SSCD with decoders of transformer, convolution, and our proposed cross branch feature fusion by 5% labeled training data. Sup-only denotes that our method only be trained by 5% labeled training data.
  • Figure 2: The architecture of our change detection network.
  • Figure 3: The framework of consistency regularization-based semi-supervised change detection method.
  • Figure 4: Detection results of different methods on WHU-CD and LEVIR-CD at the 5% labeled training ratio.