MetaAnchor: Learning to Detect Objects with Customized Anchors

Abstract

We propose a novel and flexible anchor mechanism named MetaAnchor for object detection frameworks. Unlike many previous detectors model anchors via a predefined manner, in MetaAnchor anchor functions could be dynamically generated from the arbitrary customized prior boxes. Taking advantage of weight prediction, MetaAnchor is able to work with most of the anchor-based object detection systems such as RetinaNet. Compared with the predefined anchor scheme, we empirically find that MetaAnchor is more robust to anchor settings and bounding box distributions; in addition, it also shows the potential on transfer tasks. Our experiment on COCO detection task shows that MetaAnchor consistently outperforms the counterparts in various scenarios.

Cited by in corpus (33)

• Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression
• AutoAssign: Differentiable Label Assignment for Dense Object Detection
• DetNAS: Backbone Search for Object Detection
• RepPoints: Point Set Representation for Object Detection
• MetaPruning: Meta Learning for Automatic Neural Network Channel Pruning
• Dynamic Neural Networks: A Survey
• Recent Advances in Deep Learning for Object Detection
• Detection in Crowded Scenes: One Proposal, Multiple Predictions
• Consistent Optimization for Single-Shot Object Detection
• Anchor Pruning for Object Detection
• Learning to Learn Variational Semantic Memory
• Anchor Box Optimization for Object Detection
• Multiview Detection with Feature Perspective Transformation
• MSNet: A Multilevel Instance Segmentation Network for Natural Disaster Damage Assessment in Aerial Videos
• Meta-PU: An Arbitrary-Scale Upsampling Network for Point Cloud
• LLA: Loss-aware Label Assignment for Dense Pedestrian Detection
• OTA: Optimal Transport Assignment for Object Detection
• iffDetector: Inference-aware Feature Filtering for Object Detection
• An Asymptotically Optimal Multi-Armed Bandit Algorithm and Hyperparameter Optimization
• NOH-NMS: Improving Pedestrian Detection by Nearby Objects Hallucination
• RRPN++: Guidance Towards More Accurate Scene Text Detection
• Multiple Anchor Learning for Visual Object Detection
• X-LineNet: Detecting Aircraft in Remote Sensing Images by a pair of Intersecting Line Segments
• Probabilistic Ranking-Aware Ensembles for Enhanced Object Detections
• Semi-Anchored Detector for One-Stage Object Detection
• A Study on Evaluation Standard for Automatic Crack Detection Regard the Random Fractal
• Image Synthesis via Semantic Composition
• Which to Match? Selecting Consistent GT-Proposal Assignment for Pedestrian Detection
• PointINS: Point-based Instance Segmentation
• Analysis and a Solution of Momentarily Missed Detection for Anchor-based Object Detectors
• Revisiting the Loss Weight Adjustment in Object Detection
• CFAD: Coarse-to-Fine Action Detector for Spatiotemporal Action Localization
• RMOPP: Robust Multi-Objective Post-Processing for Effective Object Detection