CS-R-FCN: Cross-supervised Learning for Large-Scale Object Detection
This work addresses the problem of reducing annotation costs for large-scale object detection in computer vision, though it appears incremental as it builds on existing two-stage detection frameworks.
The paper tackles the challenge of large-scale object detection with limited bounding-box annotations by introducing CS-R-FCN, a cross-supervised learning pipeline that combines bounding-box-level and image-level annotated data, resulting in a significant improvement in mAP compared to prior works.
Generic object detection is one of the most fundamental problems in computer vision, yet it is difficult to provide all the bounding-box-level annotations aiming at large-scale object detection for thousands of categories. In this paper, we present a novel cross-supervised learning pipeline for large-scale object detection, denoted as CS-R-FCN. First, we propose to utilize the data flow of image-level annotated images in the fully-supervised two-stage object detection framework, leading to cross-supervised learning combining bounding-box-level annotated data and image-level annotated data. Second, we introduce a semantic aggregation strategy utilizing the relationships among the cross-supervised categories to reduce the unreasonable mutual inhibition effects during the feature learning. Experimental results show that the proposed CS-R-FCN improves the mAP by a large margin compared to previous related works.