Research map / Detection and segmentation / CNN features
Searching Efficient Semantic Segmentation Architectures via Dynamic Path Selection
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes a NAS method for efficient semantic segmentation that dynamically selects promising paths within a one-shot supernet to avoid inefficient weight updates and gradient conflicts from suboptimal paths.
Problem
Neural architecture search for semantic segmentation with one-shot supernets applies uniform optimization to all candidate paths, causing inefficient weight updates and gradient conflicts due to complex multi-branch architectures and large search spaces.
Approach
Dynamic path selection during supernet training so that promising paths receive focused optimization while suboptimal paths are deprioritized, rather than uniformly updating all candidate paths in a one-shot supernet.
Experimental setup
Builds on: one-shot NAS
Closest papers by meaning
- FasterSeg: Searching for Faster Real-time Semantic Segmentation (ICLR 2020 · similarity 0.90)
- Learning Dynamic Routing for Semantic Segmentation (CVPR 2020 · similarity 0.89)
- Fast Neural Architecture Search of Compact Semantic Segmentation Models via Auxiliary Cells (CVPR 2019 · similarity 0.87)
- Diffusion-Driven Two-Stage Active Learning for Low-Budget Semantic Segmentation (NeurIPS 2025 · similarity 0.87)
- SasWOT: Real-Time Semantic Segmentation Architecture Search WithOut Training (AAAI 2024 · similarity 0.86)
- Semantic-aware SAM for Point-Prompted Instance Segmentation (CVPR 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.