DetNet: A Backbone network for Object Detection
arXiv:1804.06215
Abstract
Recent CNN based object detectors, no matter one-stage methods like YOLO, SSD, and RetinaNe or two-stage detectors like Faster R-CNN, R-FCN and FPN are usually trying to directly finetune from ImageNet pre-trained models designed for image classification. There has been little work discussing on the backbone feature extractor specifically designed for the object detection. More importantly, there are several differences between the tasks of image classification and object detection. 1. Recent object detectors like FPN and RetinaNet usually involve extra stages against the task of image classification to handle the objects with various scales. 2. Object detection not only needs to recognize the category of the object instances but also spatially locate the position. Large downsampling factor brings large valid receptive field, which is good for image classification but compromises the object location ability. Due to the gap between the image classification and object detection, we propose DetNet in this paper, which is a novel backbone network specifically designed for object detection. Moreover, DetNet includes the extra stages against traditional backbone network for image classification, while maintains high spatial resolution in deeper layers. Without any bells and whistles, state-of-the-art results have been obtained for both object detection and instance segmentation on the MSCOCO benchmark based on our DetNet~(4.8G FLOPs) backbone. The code will be released for the reproduction.
References in corpus (3)
Cited by in corpus (41)
- A Survey of Deep Learning-based Object Detection
- Object Detection in 20 Years: A Survey
- Deep Learning for UAV-based Object Detection and Tracking: A Survey
- Deep Learning for Generic Object Detection: A Survey
- GCNet: Non-local Networks Meet Squeeze-Excitation Networks and Beyond
- TPH-YOLOv5: Improved YOLOv5 Based on Transformer Prediction Head for Object Detection on Drone-captured Scenarios
- Backbones-Review: Feature Extraction Networks for Deep Learning and Deep Reinforcement Learning Approaches
- Rethinking on Multi-Stage Networks for Human Pose Estimation
- Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection
- Radio Galaxy Zoo: ClaRAN - A Deep Learning Classifier for Radio Morphologies
- Rethinking ImageNet Pre-training
- A Systematic IoU-Related Method: Beyond Simplified Regression for Better Localization
- Adversarial Example in Remote Sensing Image Recognition
- Computation Reallocation for Object Detection
- RefineDetLite: A Lightweight One-stage Object Detection Framework for CPU-only Devices
- Recent Advances in Deep Learning for Object Detection
- HRDNet: High-resolution Detection Network for Small Objects
- Solution for Large-Scale Hierarchical Object Detection Datasets with Incomplete Annotation and Data Imbalance
- PENet: Object Detection using Points Estimation in Aerial Images
- Hit-Detector: Hierarchical Trinity Architecture Search for Object Detection
- SM-NAS: Structural-to-Modular Neural Architecture Search for Object Detection
- Deep Learning in Diabetic Foot Ulcers Detection: A Comprehensive Evaluation
- ScratchDet: Training Single-Shot Object Detectors from Scratch
- DSFD: Dual Shot Face Detector
- IPG-Net: Image Pyramid Guidance Network for Small Object Detection
- Colonoscopy Polyp Detection and Classification: Dataset Creation and Comparative Evaluations
- Cheaper Pre-training Lunch: An Efficient Paradigm for Object Detection
- Augmented Parallel-Pyramid Net for Attention Guided Pose-Estimation
- TransNAS-Bench-101: Improving Transferability and Generalizability of Cross-Task Neural Architecture Search
- LIP: Local Importance-based Pooling
- A Novel and Efficient Tumor Detection Framework for Pancreatic Cancer via CT Images
- Inception Convolution with Efficient Dilation Search
- EHSOD: CAM-Guided End-to-end Hybrid-Supervised Object Detection with Cascade Refinement
- LabelEnc: A New Intermediate Supervision Method for Object Detection
- CASSOD-Net: Cascaded and Separable Structures of Dilated Convolution for Embedded Vision Systems and Applications
- G-RCN: Optimizing the Gap between Classification and Localization Tasks for Object Detection
- Decoupled Self Attention for Accurate One Stage Object Detection
- Universal-RCNN: Universal Object Detector via Transferable Graph R-CNN
- Net: Augmented Parallel-Pyramid Net for Attention Guided Pose Estimation
- Dense Multiscale Feature Fusion Pyramid Networks for Object Detection in UAV-Captured Images
- DupNet: Towards Very Tiny Quantized CNN with Improved Accuracy for Face Detection