StaticSegFormer: An Efficient High-Performance Semantic Segmentation Based on Static Structured Pruning
Not on the map. Why?
Abstract (arXiv)
Structured pruning enhances the efficiency of deep neural networks (DNNs) by eliminating groups of parameters during inference. Previous methods mostly reduce computational complexity (FLOPs), while semantic segmentation performance (mIoU) slightly drops. Accordingly, recent dynamic structured pruning methods aim at reducing the performance drop, while lowering the FLOPs even more. However, on the ADE20K and Cityscapes benchmarks, our study reveals that on a GPU platform such dynamic methods exhibit a surprisingly low frame rate far below a simple static approach, while having comparable results in mIoU and FLOPs. To address this issue, we propose a static structured pruning method for attention layers, that achieves both, a lower FLOPs and a high frame rate [fps] of the SegFormer network, the latter increased by up to 34% relative on the Cityscapes dataset, while having no mIoU performance drop at all. Our so-called StaticSegFormer method is strongest for small encoders and large images.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 31 of its 40 references that are in the map. First posted on arXiv 5 Aug 2026.
Builds on
31 of 40 references.
References · 31
- 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
- 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
- 2014Microsoft COCO: Common Objects in Contextcited by 17,608
- 2015Distilling the Knowledge in a Neural Networkcited by 10,324
- 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,442
- 2014Fully Convolutional Networks for Semantic Segmentationcited by 4,859
- 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,444
- 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,845
- 2021SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformerscited by 2,940
- 2021Masked-attention Mask Transformer for Universal Image Segmentationcited by 2,351
- 2018The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networkscited by 2,262
- 2015Learning both Weights and Connections for Efficient Neural Networkscited by 2,031
19 more
- 2014FitNets: Hints for Thin Deep Netscited by 1,653
- 2021DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsificationcited by 864
- 2019Are Sixteen Heads Really Better than One?cited by 816
- 2022EVA: Exploring the Limits of Masked Visual Representation Learning at Scalecited by 691
- 2017Learning Efficient Convolutional Networks through Network Slimmingcited by 676
- 2018Filter Pruning via Geometric Median for Deep Convolutional Neural Networks Accelerationcited by 367
- 2015Learning Deconvolution Network for Semantic Segmentationcited by 365
- 2019Data-Free Quantization Through Weight Equalization and Bias Correctioncited by 336
- 2022Vision Transformer with Deformable Attentioncited by 295
- 2021Distilling Knowledge via Knowledge Reviewcited by 288
- 2021Patch Slimming for Efficient Vision Transformerscited by 133
- 2018Dynamic Channel Pruning: Feature Boosting and Suppressioncited by 129
- 2015Fast ConvNets Using Group-wise Brain Damagecited by 82
- 2020Post-Training Piecewise Linear Quantization for Deep Neural Networkscited by 79
- 2023Dynamic Token Pruning in Plain Vision Transformers for Semantic Segmentationcited by 62
- 2019Batch-Shaping for Learning Conditional Channel Gated Networkscited by 42
- 2018Structured Binary Neural Networks for Accurate Image Classification and Semantic Segmentationcited by 37
- 2021Dynamically pruning segformer for efficient semantic segmentationcited by 13
- 2022Auto-Compressing Subset Pruning for Semantic Image Segmentationcited by 4
Built on by
Not cited yet.