mlsanityData to 14 Oct 2026

The Road to On-board Change Detection: A Lightweight Patch-Level Change Detection Network via Exploring the Potential of Pruning and Pooling

arXiv 2310.10166 ↗First posted 16 Oct 2023cs.CV cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Existing satellite remote sensing change detection (CD) methods often crop original large-scale bi-temporal image pairs into small patch pairs and then use pixel-level CD methods to fairly process all the patch pairs. However, due to the sparsity of change in large-scale satellite remote sensing images, existing pixel-level CD methods suffer from a waste of computational cost and memory resources on lots of unchanged areas, which reduces the processing efficiency of on-board platform with extremely limited computation and memory resources. To address this issue, we propose a lightweight patch-level CD network (LPCDNet) to rapidly remove lots of unchanged patch pairs in large-scale bi-temporal image pairs. This is helpful to accelerate the subsequent pixel-level CD processing stage and reduce its memory costs. In our LPCDNet, a sensitivity-guided channel pruning method is proposed to remove unimportant channels and construct the lightweight backbone network on basis of ResNet18 network. Then, the multi-layer feature compression (MLFC) module is designed to compress and fuse the multi-level feature information of bi-temporal image patch. The output of MLFC module is fed into the fully-connected decision network to generate the predicted binary label. Finally, a weighted cross-entropy loss is utilized in the training process of network to tackle the change/unchange class imbalance problem. Experiments on two CD datasets demonstrate that our LPCDNet achieves more than 1000 frames per second on an edge computation platform, i.e., NVIDIA Jetson AGX Orin, which is more than 3 times that of the existing methods without noticeable CD performance loss. In addition, our method reduces more than 60% memory costs of the subsequent pixel-level CD processing stage.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 37 of its 83 references that are in the map. First posted on arXiv 16 Oct 2023.

Builds on

37 of 83 references.

References · 37

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  3. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,364
  4. 2018MobileNetV2: Inverted Residuals and Linear Bottleneckscited by 4,795
  5. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,845
  6. 2015Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Codingcited by 2,329
  7. 2015Learning both Weights and Connections for Efficient Neural Networkscited by 2,031
  8. 2019Searching for MobileNetV3cited by 1,994
  9. 2017Deformable Convolutional Networkscited by 1,488
  10. 2016Pruning Filters for Efficient ConvNetscited by 1,143
  11. 2018ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Designcited by 1,136
  12. 2017Channel Pruning for Accelerating Very Deep Neural Networkscited by 702
25 more
  1. 2013Network In Networkcited by 605
  2. 2018Rethinking the Value of Network Pruningcited by 599
  3. 2019GhostNet: More Features from Cheap Operationscited by 410
  4. 2018Filter Pruning via Geometric Median for Deep Convolutional Neural Networks Accelerationcited by 367
  5. 2018Soft Filter Pruning for Accelerating Deep Convolutional Neural Networkscited by 315
  6. 2018Fully Convolutional Siamese Networks for Change Detectioncited by 306
  7. 2021Remote Sensing Image Change Detection with Transformerscited by 306
  8. 2016EIE: Efficient Inference Engine on Compressed Deep Neural Networkcited by 295
  9. 2020HRank: Filter Pruning using High-Rank Feature Mapcited by 292
  10. 2018Model compression via distillation and quantizationcited by 289
  11. 2020ReActNet: Towards Precise Binary Neural Network with Generalized Activation Functionscited by 184
  12. 2015Compressing Neural Networks with the Hashing Trickcited by 150
  13. 2020Strip Pooling: Rethinking Spatial Pooling for Scene Parsingcited by 114
  14. 2017Shift: A Zero FLOP, Zero Parameter Alternative to Spatial Convolutionscited by 93
  15. 2021CHIP: CHannel Independence-based Pruning for Compact Neural Networkscited by 78
  16. 2020DASNet: Dual attentive fully convolutional siamese networks for change detection of high resolution satellite imagescited by 75
  17. 2020Group Sparsity: The Hinge Between Filter Pruning and Decomposition for Network Compressioncited by 72
  18. 2013Stochastic Pooling for Regularization of Deep Convolutional Neural Networkscited by 48
  19. 2017An Entropy-based Pruning Method for CNN Compressioncited by 37
  20. 2022TINYCD: A (Not So) Deep Learning Model For Change Detectioncited by 36
  21. 2014Efficient and Accurate Approximations of Nonlinear Convolutional Networkscited by 30
  22. 2019LIP: Local Importance-based Poolingcited by 30
  23. 2021Refining activation downsampling with SoftPoolcited by 26
  24. 2020Online Ensemble Model Compression using Knowledge Distillationcited by 16
  25. 2022LSNet: Extremely Light-Weight Siamese Network For Change Detection in Remote Sensing Imagecited by 4

Built on by

Not cited yet.