ElasticFormer speeds object detection in huge images by focusing computing
ElasticFormer: Detecting Objects in HRW Shots via Elastic Computing Vision Transformer
Computer Vision and Pattern Recognition
Summary
Detecting objects in extremely large, high-resolution pictures is difficult because most of the image is empty background, which slows down and confuses standard detectors. The authors introduce ElasticFormer, a method that smartly picks which parts of the image to focus on by guessing where objects might be. This saves a lot of computer work while improving detection accuracy. Their approach adapts well even without detailed object markings, making it flexible for real-world use.
What this means in practice
- •For industrial inspection teams: Detect defects or features in extremely large images more efficiently by focusing computing on image areas likely to contain objects, reducing processing time and increasing accuracy.
- •For satellite image analysts: Improve object detection in high-resolution satellite images by dynamically allocating computing resources to areas with more objects, enabling faster and more accurate analysis.
Authors
Wenxi Li, Jingchen Huang, Chenyang Lyu, Moran Liu, Haozhe Lin, Guiguang Ding, Yuchen Guo
Abstract
Recent advances in gigapixel-level imaging have brought High-Resolution Wide shots to the forefront of research. However, these images present significant challenges: extreme sparsity of foreground, gigapixel-level resolutions and diverse target counts. This makes traditional close-up detectors inaccurate and slow as they are overwhelmed by the background. Although previous research has explored sparse backbones, their fixed sparsity patterns lack the adaptability required to handle diverse target numbers. To address this, we introduce ElasticFormer, a sparse backbone that dynamically allocates computational resources based on foreground proportion. After scoring windows based on variance, proposed ElasticSelector module will predict the foreground proportion for top-k selection. The mechanism guides the model to select target-containing windows, scaling resources in areas where objects are clustered. We introduce a novel loss function combined with the 3-phase training strategy for ElasticSelector, allowing it to function properly when bounding box annotations are missing. A WSOD study is carried on PASCAL VOC 2007 to evaluate its extensibility. Further, ElasticNet is created to verify its backbone-agnostic nature. In experiments on the PANDA gigapixel benchmark, ElasticFormer reduces backbone FLOPs by 80% while achieving a significant improvement in AP50 when compared to fixed-ratio sparse methods.