An Implementation of Faster RCNN with Study for Region Sampling
arXiv:1702.02138
Abstract
We adapted the join-training scheme of Faster RCNN framework from Caffe to TensorFlow as a baseline implementation for object detection. Our code is made publicly available. This report documents the simplifications made to the original pipeline, with justifications from ablation analysis on both PASCAL VOC 2007 and COCO 2014. We further investigated the role of non-maximal suppression (NMS) in selecting regions-of-interest (RoIs) for region classification, and found that a biased sampling toward small regions helps performance and can achieve on-par mAP to NMS-based sampling when converged sufficiently.
Technical Report, 3 pages
References in corpus (2)
Cited by in corpus (12)
- Automatic Colon Polyp Detection using Region based Deep CNN and Post Learning Approaches
- Adversarial Examples that Fool Detectors
- PixelNet: Representation of the pixels, by the pixels, and for the pixels
- Lifelong Object Detection
- ChainerCV: a Library for Deep Learning in Computer Vision
- Deep Mangoes: from fruit detection to cultivar identification in colour images of mango trees
- Fast Vehicle Detection in Aerial Imagery
- Video Object Segmentation using Tracked Object Proposals
- Intention Recognition of Pedestrians and Cyclists by 2D Pose Estimation
- Image Captioning with Integrated Bottom-Up and Multi-level Residual Top-Down Attention for Game Scene Understanding
- CPARR: Category-based Proposal Analysis for Referring Relationships
- Efficient Pipelines for Vision-Based Context Sensing