How good are detection proposals, really?
arXiv:1406.6962
Abstract
Current top performing Pascal VOC object detectors employ detection proposals to guide the search for objects thereby avoiding exhaustive sliding window search across images. Despite the popularity of detection proposals, it is unclear which trade-offs are made when using them during object detection. We provide an in depth analysis of ten object proposal methods along with four baselines regarding ground truth annotation recall (on Pascal VOC 2007 and ImageNet 2013), repeatability, and impact on DPM detector performance. Our findings show common weaknesses of existing methods, and provide insights to choose the most adequate method for different settings.
Cited by in corpus (20)
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
- Object Detection in 20 Years: A Survey
- Scalable, High-Quality Object Detection
- ME R-CNN: Multi-Expert R-CNN for Object Detection
- Tracking Randomly Moving Objects on Edge Box Proposals
- Beyond Local Search: Tracking Objects Everywhere with Instance-Specific Proposals
- DeepBox: Learning Objectness with Convolutional Networks
- Deep Neural Networks for Marine Debris Detection in Sonar Images
- A pooling based scene text proposal technique for scene text reading in the wild
- Actionness Estimation Using Hybrid Fully Convolutional Networks
- Self-taught Object Localization with Deep Networks
- Weakly Supervised Representation Learning for Unsynchronized Audio-Visual Events
- Exploit Bounding Box Annotations for Multi-label Object Recognition
- Semantic Hierarchical Priors for Intrinsic Image Decomposition
- Diagnosing State-Of-The-Art Object Proposal Methods
- Object Proposal Generation using Two-Stage Cascade SVMs
- Cascaded Sparse Spatial Bins for Efficient and Effective Generic Object Detection
- Evaluation of Object Detection Proposals Under Condition Variations
- Active Object Localization with Deep Reinforcement Learning
- Model-based active learning to detect isometric deformable objects in the wild with deep architectures