LeMoRe: Learn More Details for Lightweight Semantic Segmentation

2025/05/29 by Mian Muhammad Naeem Abid, Nancy Mehta, Abid, Mian Muhammad Naeem +5
Computer Science · #Advanced Neural Network Applications #Computer Vision and Pattern Recognition (cs.CV) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Multimodal Machine Learning Applications

paper · pdf · doi:10.48550/arxiv.2505.23093

openalex publication_date 2025/05/29 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Lightweight semantic segmentation is essential for many downstream vision tasks. Unfortunately, existing methods often struggle to balance efficiency and performance due to the complexity of feature modeling. Many of these existing approaches are constrained by rigid architectures and implicit representation learning, often characterized by parameter-heavy designs and a reliance on computationally intensive Vision Transformer-based frameworks. In this work, we introduce an efficient paradigm by synergizing explicit and implicit modeling to balance computational efficiency with representational fidelity. Our method combines well-defined Cartesian directions with explicitly modeled views and implicitly inferred intermediate representations, efficiently capturing global dependencies through a nested attention mechanism. Extensive experiments on challenging datasets, including ADE20K, CityScapes, Pascal Context, and COCO-Stuff, demonstrate that LeMoRe strikes an effective balance between performance and efficiency.

Citations

• PEM: Prototype-based Efficient MaskFormer for Image Segmentation
• MetaSeg: MetaFormer-based Global Contexts-aware Network for Efficient Semantic Segmentation
• U-MixFormer: UNet-like Transformer with Mix-Attention for Efficient Semantic Segmentation
• SegNeXt: Rethinking Convolutional Attention Design for Semantic Segmentation
• EfficientFormer: Vision Transformers at MobileNet Speed
• TopFormer: Token Pyramid Transformer for Mobile Semantic Segmentation
• MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer
• UNetFormer: A UNet-like transformer for efficient semantic segmentation of remote sensing urban scene imagery
• ConvMLP: Hierarchical Convolutional MLPs for Vision
• HR-NAS: Searching Efficient High-Resolution Neural Architectures with Lightweight Transformers
• SegFormer: Simple and Efficient Design for Semantic Segmentation with\n Transformers
• LeViT: a Vision Transformer in ConvNet's Clothing for Faster Inference
• Training data-efficient image transformers & distillation through attention
• A Systematic Review and Meta-Analysis of Takeover Performance During Conditionally Automated Driving
• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
• GhostNet: More Features from Cheap Operations
• GhostNet: More Features From Cheap Operations
• EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
• Searching for MobileNetV3
• Panoptic Feature Pyramid Networks
• ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design
• Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation
• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• COCO-Stuff: Thing and Stuff Classes in Context
• Pyramid Scene Parsing Network
• ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation
• Cityscapes dataset for semantic urban scene understanding
• Deep Residual Learning for Image Recognition
• Fully Convolutional Networks for Semantic Segmentation

Related

• Stronger, Fewer, & Superior: Harnessing Vision Foundation Models for Domain Generalized Semantic Segmentation
• Low-Resolution Self-Attention for Semantic Segmentation
• PP-LiteSeg: A Superior Real-Time Semantic Segmentation Model
• Revisiting Efficient Semantic Segmentation: Learning Offsets for Better Spatial and Class Feature Alignment
• CLIP for Lightweight Semantic Segmentation