New EMCStereo method enhances depth estimation for thin structures

Researchers have developed EMCStereo, a novel stereo matching method designed to improve depth estimation for thin structures like tree branches. The method integrates three lightweight attention modules—Efficient Multi-scale Attention (EMA), a Multi-Scale Fusion block (MSFblock), and Coordinate Attention (CoordAtt)—into a PSMNet-style backbone, resulting in a model that is only slightly larger and has minimal inference time overhead. To evaluate EMCStereo, a synthetic dataset called VirtualTree was created using Unreal Engine 5, featuring precise disparity labels for thin branches. The model achieved strong performance on VirtualTree and several established benchmarks, including KITTI 2012, KITTI 2015, ETH3D, and Middlebury. AI

IMPACT Improves depth estimation for challenging thin structures, potentially benefiting robotics and autonomous systems.

RANK_REASON Academic paper detailing a new method and dataset for computer vision. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• Efficient Multi-scale Attention
• EMCStereo
• ETH3D
• KITTI 2012
• KITTI 2015
• Middlebury
• Multi-Scale Fusion block
• PSMNet
• Unreal Engine 5
• VirtualTree
• ZED Mini

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →