Scaled-YOLOv4: Scaling Cross Stage Partial Network

Abstract

We show that the YOLOv4 object detection neural network based on the CSP approach, scales both up and down and is applicable to small and large networks while maintaining optimal speed and accuracy. We propose a network scaling approach that modifies not only the depth, width, resolution, but also structure of the network. YOLOv4-large model achieves state-of-the-art results: 55.5% AP (73.4% AP50) for the MS COCO dataset at a speed of ~16 FPS on Tesla V100, while with the test time augmentation, YOLOv4-large achieves 56.0% AP (73.3 AP50). To the best of our knowledge, this is currently the highest accuracy on the COCO dataset among any published work. The YOLOv4-tiny model achieves 22.0% AP (42.0% AP50) at a speed of 443 FPS on RTX 2080Ti, while by using TensorRT, batch size = 4 and FP16-precision the YOLOv4-tiny achieves 1774 FPS.

Added references. Corrected typos. The new results are slightly better

References in corpus (12)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• YOLOv4: Optimal Speed and Accuracy of Object Detection
• YOLOv3: An Incremental Improvement
• Wide Residual Networks
• Learning Spatial Fusion for Single-Shot Object Detection
• PP-YOLO: An Effective and Efficient Implementation of Object Detector
• CornerNet-Lite: Efficient Keypoint Based Object Detection
• DetectoRS: Detecting Objects with Recursive Feature Pyramid and Switchable Atrous Convolution
• SpineNet: Learning Scale-Permuted Backbone for Recognition and Localization
• RDSNet: A New Deep Architecture for Reciprocal Object Detection and Instance Segmentation
• BorderDet: Border Feature for Dense Object Detection

Cited by in corpus (12)

• YOLOX: Exceeding YOLO Series in 2021
• You Only Learn One Representation: Unified Network for Multiple Tasks
• You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection
• Probabilistic two-stage detection
• Sparse R-CNN: End-to-End Object Detection with Learnable Proposals
• GIAOTracker: A comprehensive framework for MCMOT with global information and optimizing strategies in VisDrone 2021
• Rock Hunting With Martian Machine Vision
• 2nd Place Solution for Waymo Open Dataset Challenge -- Real-time 2D Object Detection
• DetectorGuard: Provably Securing Object Detectors against Localized Patch Hiding Attacks
• Disentangle Your Dense Object Detector
• Pulmonary embolism identification in computerized tomography pulmonary angiography scans with deep learning technologies in COVID-19 patients
• Scorpion detection and classification systems based on computer vision and deep learning for health security purposes