Abstract:Over the past years, YOLOs have emerged as the predominant paradigm in the field of real-time object detection owing to their effective balance between computational cost and detection performance. Researchers have explored the architectural designs, optimization objectives, data augmentation strategies, and others for YOLOs, achieving notable progress. However, the reliance on the non-maximum suppression (NMS) for post-processing hampers the end-to-end deployment of YOLOs and adversely impacts the inference latency. Besides, the design of various components in YOLOs lacks the comprehensive and thorough inspection, resulting in noticeable computational redundancy and limiting the model's capability. It renders the suboptimal efficiency, along with considerable potential for performance improvements. In this work, we aim to further advance the performance-efficiency boundary of YOLOs from both the post-processing and model architecture. To this end, we first present the consistent dual assignments for NMS-free training of YOLOs, which brings competitive performance and low inference latency simultaneously. Moreover, we introduce the holistic efficiency-accuracy driven model design strategy for YOLOs. We comprehensively optimize various components of YOLOs from both efficiency and accuracy perspectives, which greatly reduces the computational overhead and enhances the capability. The outcome of our effort is a new generation of YOLO series for real-time end-to-end object detection, dubbed YOLOv10. Extensive experiments show that YOLOv10 achieves state-of-the-art performance and efficiency across various model scales. For example, our YOLOv10-S is 1.8$\times$ faster than RT-DETR-R18 under the similar AP on COCO, meanwhile enjoying 2.8$\times$ smaller number of parameters and FLOPs. Compared with YOLOv9-C, YOLOv10-B has 46\% less latency and 25\% fewer parameters for the same performance.

YOLO-LITE: A Real-Time Object Detection Algorithm Optimized for Non-GPU Computers.

Mixed YOLOv3-LITE: A Lightweight Real-Time Object Detection Method

YOLObile: Real-Time Object Detection on Mobile Devices via Compression-Compilation Co-Design

SlimYOLOv4: Lightweight Object Detector Based on YOLOv4

Real-time Object Detector for Low-End Devices

Lite-YOLOv3: a real-time object detector based on multi-scale slice depthwise convolution and lightweight attention mechanism

YOLOv4-dense: A smaller and faster YOLOv4 for real-time edge-device based object detection in traffic scene

TRC‐YOLO: A Real‐time Detection Method for Lightweight Targets Based on Mobile Devices

YOLO9000: Better, Faster, Stronger

Lightweight Real-Time Object Detection Based on Deep Learning.

YOLOv10: Real-Time End-to-End Object Detection

Smart-YOLO: A Light-Weight Real-time Object Detection Network

TF-YOLO: an Improved Incremental Network for Real-Time Object Detection

Real-time object detection method based on YOLOv5 and efficient mobile network

A Real-Time Object Detection Algorithm for Video

LeYOLO, New Scalable and Efficient CNN Architecture for Object Detection

Making You Only Look Once Faster: Toward Real-Time Intelligent Transportation Detection

Brisk-Yolo: A Lightweight Object Detection Algorithm for Edge Devices.

MobileYOLO: Real-Time Object Detection Algorithm in Autonomous Driving Scenarios

GL-YOLO-Lite: A Novel Lightweight Fallen Person Detection Model

Real-time Detection Algorithm for Non-Motorized Vehicles Based on D-YOLO Model.