Real-Time Object Detection

Description from the archive archive 2025-07-28.

Benchmarks archive 2025-07-28

8 leaderboard tables shown for this task, 8 with rows (a “benchmark” on this site is a table with at least one row, as on /sota), ordered by row count. “Best model” is the first row in the archive's own order at snapshot; nothing is re-ranked here and metric direction is not recorded in the archive. PwC's Trend sparklines are not in the archive, so that column is omitted.

Syntology column: samples harvested from the paper's repositories and executed on synthesized fixtures; “ran” is not a correctness claim and does not order the table. A dash means no Syntology record for that paper, not a recorded non-run. Read from the graph 2026-09-24.

Libraries

Not in the archive: the export carries no per-task library table, so there is nothing to show at snapshot 2025-07-28.

Datasets archive 2025-07-28

9 datasets whose archive record lists this task, ordered by the archive's paper count.

Subtasks archive 2025-07-28

No subtask under this task in the archive's task tree.

Parent tasks archive 2025-07-28

Most implemented papers archive 2025-07-28

30 shown of 134 papers with code (259 tagged with this task in all), ordered by repositories listed in the archive, not by stars (the archive holds no stars, so PwC's “Social” and “Latest” sorts cannot be reproduced). Papers without a page here are shown as plain text.

• YOLOv3: An Incremental Improvement
  8 Apr 2018 311 repositories listed Syntology ran 18 of 124 samples · 106 unverified · 19 pointer-only (licence)
  At 320x320 YOLOv3 runs in 22 ms at 28.
• Focal Loss for Dense Object Detection
  7 Aug 2017 234 repositories listed Syntology ran 11 of 11 samples · 0 unverified · 6 pointer-only (licence)
  Our novel Focal Loss focuses training on a sparse set of hard examples and prevents the vast number of easy negatives from overwhelming the detector during training.
• YOLO9000: Better, Faster, Stronger
  25 Dec 2016 231 repositories listed Syntology ran 16 of 60 samples · 44 unverified · 22 pointer-only (licence)
  On the 156 classes not in COCO, YOLO9000 gets 16.
• YOLOv4: Optimal Speed and Accuracy of Object Detection
  23 Apr 2020 223 repositories listed Syntology ran 24 of 184 samples · 160 unverified · 8 pointer-only (licence)
  There are a huge number of features which are said to improve Convolutional Neural Network (CNN) accuracy.
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
  4 Jun 2015 196 repositories listed Syntology ran 59 of 124 samples · 65 unverified · 42 pointer-only (licence)
  In this work, we introduce a Region Proposal Network (RPN) that shares full-image convolutional features with the detection network, thus enabling nearly cost-free region proposals.
• Mask R-CNN
  20 Mar 2017 179 repositories listed Syntology ran 42 of 140 samples · 98 unverified · 23 pointer-only (licence)
  Our approach efficiently detects objects in an image while simultaneously generating a high-quality segmentation mask for each instance.
• You Only Look Once: Unified, Real-Time Object Detection
  8 Jun 2015 144 repositories listed Syntology ran 80 of 148 samples · 68 unverified · 98 pointer-only (licence)
  A single neural network predicts bounding boxes and class probabilities directly from full images in one evaluation.
• CSPNet: A New Backbone that can Enhance Learning Capability of CNN
  27 Nov 2019 123 repositories listed
  Neural networks have enabled state-of-the-art approaches to achieve incredible results on computer vision tasks such as object detection.
• Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
  25 Mar 2021 80 repositories listed Syntology ran 108 of 207 samples · 99 unverified · 43 pointer-only (licence)
  This paper presents a new vision Transformer, called Swin Transformer, that capably serves as a general-purpose backbone for computer vision.
• Objects as Points
  16 Apr 2019 76 repositories listed Syntology ran 10 of 130 samples · 120 unverified
  We model an object as a single point --- the center point of its bounding box.
• EfficientDet: Scalable and Efficient Object Detection
  20 Nov 2019 64 repositories listed Syntology ran 11 of 70 samples · 59 unverified · 3 pointer-only (licence)
  Model efficiency has become increasingly important in computer vision.
• A ConvNet for the 2020s
  10 Jan 2022 54 repositories listed Syntology ran 54 of 80 samples · 26 unverified · 11 pointer-only (licence)
  The "Roaring 20s" of visual recognition began with the introduction of Vision Transformers (ViTs), which quickly superseded ConvNets as the state-of-the-art image classification model.
• R-FCN: Object Detection via Region-based Fully Convolutional Networks
  20 May 2016 48 repositories listed Syntology ran 1 of 11 samples · 10 unverified
  In contrast to previous region-based detectors such as Fast/Faster R-CNN that apply a costly per-region subnetwork hundreds of times, our region-based detector is fully convolutional with almost all computation shared…
• YOLOX: Exceeding YOLO Series in 2021
  18 Jul 2021 42 repositories listed Syntology ran 1 of 23 samples · 22 unverified
  In this report, we present some experienced improvements to YOLO series, forming a new high-performance detector -- YOLOX.
• Scaled-YOLOv4: Scaling Cross Stage Partial Network
  16 Nov 2020 41 repositories listed Syntology ran 1 of 1 samples · 0 unverified · 1 pointer-only (licence)
  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.
• End-to-End Object Detection with Transformers
  26 May 2020 37 repositories listed Syntology ran 59 of 92 samples · 33 unverified · 19 pointer-only (licence)
  We present a new method that views object detection as a direct set prediction problem.
• MnasNet: Platform-Aware Neural Architecture Search for Mobile
  31 Jul 2018 29 repositories listed Syntology ran 2 of 6 samples · 4 unverified · 2 pointer-only (licence)
  In this paper, we propose an automated mobile neural architecture search (MNAS) approach, which explicitly incorporate model latency into the main objective so that the search can identify a model that achieves a good…
• HarDNet: A Low Memory Traffic Network
  3 Sep 2019 24 repositories listed
  We propose a Harmonic Densely Connected Network to achieve high efficiency in terms of both low MACs and memory traffic.
• YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors
  6 Jul 2022 21 repositories listed Syntology ran 1 of 11 samples · 10 unverified
  YOLOv7 surpasses all known object detectors in both speed and accuracy in the range from 5 FPS to 160 FPS and has the highest accuracy 56.
• Deformable DETR: Deformable Transformers for End-to-End Object Detection
  8 Oct 2020 20 repositories listed Syntology ran 29 of 55 samples · 26 unverified · 21 pointer-only (licence)
  DETR has been recently proposed to eliminate the need for many hand-designed components in object detection while demonstrating good performance.
• DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection
  7 Mar 2022 16 repositories listed Syntology ran 7 of 15 samples · 8 unverified · 5 pointer-only (licence)
  Compared to other models on the leaderboard, DINO significantly reduces its model size and pre-training data size while achieving better results.
• RTMDet: An Empirical Study of Designing Real-Time Object Detectors
  14 Dec 2022 14 repositories listed Syntology ran 3 of 20 samples · 17 unverified
  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.
• SpineNet: Learning Scale-Permuted Backbone for Recognition and Localization
  10 Dec 2019 13 repositories listed
  We propose SpineNet, a backbone with scale-permuted intermediate features and cross-scale connections that is learned on an object detection task by Neural Architecture Search.
• An Energy and GPU-Computation Efficient Backbone Network for Real-Time Object Detection
  22 Apr 2019 12 repositories listed
  As DenseNet conserves intermediate features with diverse receptive fields by aggregating them with dense connection, it shows good performance on the object detection task.
• SqueezeDet: Unified, Small, Low Power Fully Convolutional Neural Networks for Real-Time Object Detection for Autonomous Driving
  4 Dec 2016 12 repositories listed Syntology ran 0 of 30 samples · 30 unverified
  In addition to requiring high accuracy to ensure safety, object detection for autonomous driving also requires real-time inference speed to guarantee prompt vehicle control, as well as small model size and energy…
• ESPNetv2: A Light-weight, Power Efficient, and General Purpose Convolutional Neural Network
  28 Nov 2018 10 repositories listed Syntology ran 2 of 4 samples · 2 unverified
  Compared to YOLOv2 on the MS-COCO object detection, ESPNetv2 delivers 4.
• DETRs Beat YOLOs on Real-time Object Detection
  17 Apr 2023 9 repositories listed Syntology ran 2 of 4 samples · 2 unverified
  Our RT-DETR-R50 / R101 achieves 53.
• You Only Learn One Representation: Unified Network for Multiple Tasks
  10 May 2021 9 repositories listed
  In this paper, we propose a unified network to encode implicit knowledge and explicit knowledge together, just like the human brain can learn knowledge from normal learning as well as subconsciousness learning.
• Pelee: A Real-Time Object Detection System on Mobile Devices
  18 Apr 2018 9 repositories listed
  In this study, we propose an efficient architecture named PeleeNet, which is built with conventional convolution instead.
• PVANet: Lightweight Deep Neural Networks for Real-time Object Detection
  23 Nov 2016 9 repositories listed
  In object detection, reducing computational cost is as important as improving accuracy for most practical usages.

Syntology lines on 23 of the papers shown; no Syntology record for the others (a paper without an arXiv id cannot be joined to the graph, and absence from the graph layer is not a recorded non-run). “Ran” means the sample executed on a synthesized fixture, not that the paper's result was reproduced. Read from the graph 2026-09-24.