Scale-Aware Trident Networks for Object Detection
arXiv:1901.01892
Abstract
Scale variation is one of the key challenges in object detection. In this work, we first present a controlled experiment to investigate the effect of receptive fields for scale variation in object detection. Based on the findings from the exploration experiments, we propose a novel Trident Network (TridentNet) aiming to generate scale-specific feature maps with a uniform representational power. We construct a parallel multi-branch architecture in which each branch shares the same transformation parameters but with different receptive fields. Then, we adopt a scale-aware training scheme to specialize each branch by sampling object instances of proper scales for training. As a bonus, a fast approximation version of TridentNet could achieve significant improvements without any additional parameters and computational cost compared with the vanilla detector. On the COCO dataset, our TridentNet with ResNet-101 backbone achieves state-of-the-art single-model results of 48.4 mAP. Codes are available at https://git.io/fj5vR.
arXiv commentICCV 2019 camera ready
References in corpus (10)
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
- Rethinking Atrous Convolution for Semantic Image Segmentation
- YOLOv3: An Incremental Improvement
- R-FCN: Object Detection via Region-based Fully Convolutional Networks
- MXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems
- DSSD : Deconvolutional Single Shot Detector
- Understanding the Effective Receptive Field in Deep Convolutional Neural Networks
- Beyond Skip Connections: Top-Down Modulation for Object Detection
- Light-Head R-CNN: In Defense of Two-Stage Object Detector
- MegDet: A Large Mini-Batch Object Detector
Cited by in corpus (30)
- MMDetection: Open MMLab Detection Toolbox and Benchmark
- Object Detection in 20 Years: A Survey
- Learning Spatial Fusion for Single-Shot Object Detection
- Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection
- CenterNet: Keypoint Triplets for Object Detection
- From SLAM to Situational Awareness: Challenges and Survey
- Cascade RetinaNet: Maintaining Consistency for Single-Stage Object Detection
- AugFPN: Improving Multi-scale Feature Learning for Object Detection
- CentripetalNet: Pursuing High-quality Keypoint Pairs for Object Detection
- Dynamic Scale Training for Object Detection
- Utilizing the Instability in Weakly Supervised Object Detection
- SimpleDet: A Simple and Versatile Distributed Framework for Object Detection and Instance Recognition
- SM-NAS: Structural-to-Modular Neural Architecture Search for Object Detection
- Revisiting the Sibling Head in Object Detector
- NETNet: Neighbor Erasing and Transferring Network for Better Single Shot Object Detection
- SPCNet:Spatial Preserve and Content-aware Network for Human Pose Estimation
- Towards large-scale, automated, accurate detection of CCTV camera objects using computer vision. Applications and implications for privacy, safety, and cybersecurity. (Preprint)
- Learning from Noisy Anchors for One-stage Object Detection
- Large-Scale Object Detection in the Wild from Imbalanced Multi-Labels
- SaccadeNet: A Fast and Accurate Object Detector
- Augmented Parallel-Pyramid Net for Attention Guided Pose-Estimation
- LIP: Local Importance-based Pooling
- Dually Supervised Feature Pyramid for Object Detection and Segmentation
- Instance Scale Normalization for image understanding
- Representation Sharing for Fast Object Detector Search and Beyond
- Object as Distribution
- Investigations of the Influences of a CNN's Receptive Field on Segmentation of Subnuclei of Bilateral Amygdalae
- Tiny Obstacle Discovery by Occlusion-Aware Multilayer Regression
- Fast Object Detection with Latticed Multi-Scale Feature Fusion
- Net: Augmented Parallel-Pyramid Net for Attention Guided Pose Estimation