RTMDet: An Empirical Study of Designing Real-Time Object Detectors
read the original abstract
In this paper, we aim to design an efficient real-time object detector that exceeds the YOLO series and is easily extensible for many object recognition tasks such as instance segmentation and rotated object detection. To obtain a more efficient model architecture, we explore an architecture that has compatible capacities in the backbone and neck, constructed by a basic building block that consists of large-kernel depth-wise convolutions. We further introduce soft labels when calculating matching costs in the dynamic label assignment to improve accuracy. Together with better training techniques, the resulting object detector, named RTMDet, achieves 52.8% AP on COCO with 300+ FPS on an NVIDIA 3090 GPU, outperforming the current mainstream industrial detectors. RTMDet achieves the best parameter-accuracy trade-off with tiny/small/medium/large/extra-large model sizes for various application scenarios, and obtains new state-of-the-art performance on real-time instance segmentation and rotated object detection. We hope the experimental results can provide new insights into designing versatile real-time object detectors for many object recognition tasks. Code and models are released at https://github.com/open-mmlab/mmdetection/tree/3.x/configs/rtmdet.
This paper has not been read by Pith yet.
Forward citations
Cited by 20 Pith papers
-
Geometrically Constrained Stenosis Editing in Coronary Angiography via Entropic Optimal Transport
OT-Bridge Editor uses geometrically constrained entropic optimal transport to synthesize CAG images with precise stenosis, improving downstream detection by 27.8% on ARCADE and 23.0% on a multi-center dataset.
-
SignMAE: Segmentation-Driven Self-Supervised Learning for Sign Language Recognition
SignMAE uses segmentation-driven masking in a mask-and-reconstruct self-supervised task to learn fine-grained sign representations, achieving state-of-the-art accuracy on WLASL, NMFs-CSL, and Slovo with fewer frames a...
-
KAConvNet: Kolmogorov-Arnold Convolutional Networks for Vision Recognition
KAConvNet introduces a Kolmogorov-Arnold Convolutional Layer to build networks competitive with ViTs and CNNs while offering stronger theoretical interpretability.
-
LEVIRDet: A Million-Scale 159-Category Dataset and Foundation Model for Universal Remote Sensing Object Detection
LEVIRDet-159 is a 159-category remote sensing detection dataset with 2.56M boxes exceeding prior scales; LEVIRDetNet achieves SOTA zero-shot performance on 9 external benchmarks with 5.02 mAP average improvement.
-
MORI-Seg: Learning Morphological Geometry for Instance Segmentation without Instance Annotations
MORI-Seg learns morphology-aware geometric representations from semantic masks to enable instance segmentation without instance-level annotations.
-
SLIP-RS: Structured-Attribute Language-Image Pre-Training for Remote Sensing Object Detection
SLIP-RS introduces a Structured-Attribute Decoupling Paradigm with contrastive learning and a conformal reliability engine to create a 15M-attribute dataset for remote sensing pre-training.
-
Geometrically Constrained Stenosis Editing in Coronary Angiography via Entropic Optimal Transport
OT-Bridge Editor reframes localized image editing as a constrained entropic optimal transport problem to generate synthetic coronary angiograms that boost downstream stenosis detection by 27.8% on ARCADE and 23.0% on ...
-
High-Fidelity Single-Image Head Modeling with Industry-Grade Topology
A single-image head reconstruction method uses coarse-to-fine optimization with normal consistency, landmarks, and geometry-aware constraints on curvature and conformality to produce meshes with industry-grade topolog...
-
KD-Judge: A Knowledge-Driven Automated Judge Framework for Functional Fitness Movements on Edge Devices
KD-Judge structures fitness rules via LLM retrieval and chain-of-thought, then uses pose-guided kinematics for rule-based rep validation with caching for efficient edge deployment, achieving RTF < 1 and speedups up to...
-
HMR-Net: Hierarchical Modular Routing for Cross-Domain Object Detection in Aerial Images
HMR-Net introduces hierarchical routing with global dataset-level and local scene-level modularity plus conditional experts to improve cross-domain aerial object detection and enable novel category recognition without...
-
SpikeDet: Better Firing Patterns for Accurate and Energy-Efficient Object Detection with Spiking Neural Networks
SpikeDet reaches 52.2% AP on COCO 2017 with spiking networks by optimizing firing patterns via MDSNet and SMFM, using half the energy of prior SNN detectors.
-
Learned Non-Maximum Suppression for 3D Object Detection
Two learned post-processing modules (D2D-Rescore and GossipNet3D) improve mAP, NDS, and true-positive metrics over CircleNMS on nuScenes, especially for small and rare classes.
-
FDDet: Achieving Data-Efficient Food Defect Detection Under Real-World Scenarios
FDDet is a semi-supervised object detection framework with BBoxMixUp and CGPC that outperforms standard detectors on the new FDD-48 food defect dataset under data-limited real-world conditions.
-
OSS: Open Suturing Skills Vision-Based Assessment Challenge 2024-2025
The OSS Challenge provides benchmarks showing spatiotemporal video models excel at open suturing skill classification and OSATS scoring but struggle with keypoint tracking under occlusion.
-
Towards Universal Physical Adversarial Attacks via a Joint Multi-Objective and Multi-Model Optimization Framework
JMOF is a new optimization framework for physical adversarial attacks that improves cross-model transferability and enables simultaneous attacks on multiple vision tasks such as object detection and semantic segmentation.
-
Hausdorff Distance Matching with Adaptive Query Denoising for Rotated Detection Transformer
Hausdorff distance-based matching and adaptive query denoising improve Rotated DETR, yielding +4.18 to +4.99 AP50 gains on DOTA-v2.0, DOTA-v1.5, and DIOR-R with ResNet-50.
-
Mitosis Detection in the Wild: Multi-Tumor and Context-Aware Generalization in the MIDOG 2025 Challenge
MIDOG 2025 challenge shows top mitosis detection F1 of 0.740 and atypical figure balanced accuracy of 0.908 across diverse tumors, with clear drops in challenging regions and tumor-type variation.
-
Structure-Guided Mixed Masked Pretraining and Spatial Continuity Regularization for Printed Circuit Board Defect Detection
A new PCB defect detection method using structure-guided masked pretraining and spatial continuity regularization achieves 85.5% mAP0.5 on the DsPCBSD+ dataset.
-
4th Workshop on Maritime Computer Vision (MaCVi): Challenge Overview
The report overviews five maritime computer vision benchmark challenges, their datasets, protocols, quantitative results, and top team approaches from the MaCVi 2026 workshop.
-
AIM 2025 Rip Current Segmentation (RipSeg) Challenge Report
The AIM 2025 RipSeg Challenge report presents results from five submissions on single-class instance segmentation of rip currents, highlighting deep learning and domain adaptation techniques on a diverse beach dataset.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.