ddlee-cn / Obj_det_progress_tracker
Licence: gpl-3.0
Object Detection Progress Tracker
Stars: ✭ 119
Projects that are alternatives of or similar to Obj det progress tracker
Feature intertwiner
Codebase of the paper "Feature Intertwiner for Object Detection", ICLR 2019
Stars: ✭ 111 (-6.72%)
Mutual labels: object-detection
Tfjs Tiny Yolov2
Tiny YOLO v2 object detection with tensorflow.js.
Stars: ✭ 115 (-3.36%)
Mutual labels: object-detection
Yolo V3 Iou
YOLO3 动漫人脸检测 (Based on keras and tensorflow) 2019-1-19
Stars: ✭ 116 (-2.52%)
Mutual labels: object-detection
Tensorflow Object Detection Tutorial
The purpose of this tutorial is to learn how to install and prepare TensorFlow framework to train your own convolutional neural network object detection classifier for multiple objects, starting from scratch
Stars: ✭ 113 (-5.04%)
Mutual labels: object-detection
Colab Mask Rcnn
How to run Object Detection and Segmentation on a Video Fast for Free
Stars: ✭ 114 (-4.2%)
Mutual labels: object-detection
Satellite Image Object Detection
YOLO/YOLOv2 inspired deep network for object detection on satellite images (Tensorflow, Numpy, Pandas).
Stars: ✭ 115 (-3.36%)
Mutual labels: object-detection
Bag Of Visual Words Python
Implementing Bag of Visual words approach for object classification and detection
Stars: ✭ 109 (-8.4%)
Mutual labels: object-detection
Objectdetection
Some experiments with object detection in PyTorch
Stars: ✭ 117 (-1.68%)
Mutual labels: object-detection
Yolo mark
GUI for marking bounded boxes of objects in images for training neural network Yolo v3 and v2
Stars: ✭ 1,624 (+1264.71%)
Mutual labels: object-detection
Hrnet Maskrcnn Benchmark
Object detection with multi-level representations generated from deep high-resolution representation learning (HRNetV2h).
Stars: ✭ 116 (-2.52%)
Mutual labels: object-detection
Pytorch cpp
Deep Learning sample programs using PyTorch in C++
Stars: ✭ 114 (-4.2%)
Mutual labels: object-detection
Id Card Detector
💳 Detecting the National Identification Cards with Deep Learning (Faster R-CNN)
Stars: ✭ 114 (-4.2%)
Mutual labels: object-detection
Varifocalnet
VarifocalNet: An IoU-aware Dense Object Detector
Stars: ✭ 115 (-3.36%)
Mutual labels: object-detection
Mobilenet Yolo
MobileNetV2-YoloV3-Nano: 0.5BFlops 3MB HUAWEI P40: 6ms/img, YoloFace-500k:0.1Bflops 420KB🔥🔥🔥
Stars: ✭ 1,566 (+1215.97%)
Mutual labels: object-detection
Sightseq
Computer vision tools for fairseq, containing PyTorch implementation of text recognition and object detection
Stars: ✭ 116 (-2.52%)
Mutual labels: object-detection
Tensorflow Yolov4 Tflite
YOLOv4, YOLOv4-tiny, YOLOv3, YOLOv3-tiny Implemented in Tensorflow 2.0, Android. Convert YOLO v4 .weights tensorflow, tensorrt and tflite
Stars: ✭ 1,881 (+1480.67%)
Mutual labels: object-detection
Robust Physical Attack
Physical adversarial attack for fooling the Faster R-CNN object detector
Stars: ✭ 115 (-3.36%)
Mutual labels: object-detection
Yolo3d Yolov4 Pytorch
YOLO3D: End-to-end real-time 3D Oriented Object Bounding Box Detection from LiDAR Point Cloud (ECCV 2018)
Stars: ✭ 119 (+0%)
Mutual labels: object-detection
Opencv Detection Models
OpenCV trained detection models: Haar, HOG and other
Stars: ✭ 116 (-2.52%)
Mutual labels: object-detection
Bmaskr Cnn
Boundary-preserving Mask R-CNN (ECCV 2020)
Stars: ✭ 116 (-2.52%)
Mutual labels: object-detection
Object Detection Progress Tracker
[2019.03 Update]This repo collected stat-of-the-art results on Object Detection task before 2018.03. Since I've changed my research interest, I will not update this page anymore. A few recommended places for quickly learning about the area are:
- Object Detection Paper List(Also a great review!)
- Paper With Code Leaderboards of COCO, Pascal, etc
PS: I still recommend that you should take notice of tricks the author used and FPS as well as their hardware platforms when evaluating a method. Happy Deep Learning.
Last few years have seen great progress made in the domain of Object Detection. With powerful classification networks as backbone, more and more well-designed detection heads have been proposed to handle the dilemma of recognition and localization. This repo serves as a tracker of these progress.
2-stage
| Model | Backbone | Tricks | VOC | COCO | FPS | Paper | Date | Note | Code |
|---|---|---|---|---|---|---|---|---|---|
| R-CNN | VGG16 | - | 58.5 | - | - | Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation | - | CVPR2014 Oral | MATLAB |
| Fast R-CNN | VGG16 | - | 70.0 | 19.7 | - | Fast R-CNN | 15.04 | CVPR2014 Oral | caffe |
| Faster R-CNN | VGG16 | - | 73.2 | - | - | Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks | 15.06 | NIPS2015 | MATLAB |
| Faster R-CNN+++ | ResNet-101-C4 | 1,2,3 | 73.8 | 34.9 | - | Deep Residual Learning for Image Recognition | 15.12 | CVPR2016 Best Paper | |
| R-FCN | ResNet-101 | 4 | 80.5 | 29.9 | 6 | R-FCN: Object Detection via Region-based Fully Convoluational Networks | 16.05 | NIPS2016 | caffe |
| Deformable Conv | Aligned-Inception-ResNet | 1,3 | - | 37.5 | - | Deformable Convolutional Networks | 17.03 | ICCV2017 Oral | MXNet |
| Faster R-CNN w FPN | ResNet-101-FPN | - | - | 36.2 | - | Feature Pyramid Networks for Object Detection | 16.12 | CVPR2017 Poster | caffe2 |
| Faster R-CNN by G-RMI | Inception-ResNet-v2 | - | - | 34.7 | - | Speed/accuracy Trade-offs for Modern Convolutional object detectors | - | COCO206 winner | TesnsorFlow |
| Mask R-CNN | ResNeXt | 6 | - | 39.8 | - | Mask R-CNN | 17.03 | ICCV2017 Best Paper | caffe2 |
| Light-haed R-CNN | ResNet-101-FPN | 4,6 | - | 41.5 | - | Light-Head R-CNN: In Defense of Two-stage Object Detector | 17.11 | - | |
| Light-haed R-CNN | Xception | - | - | 30.7 | 102 | Light-Head R-CNN: In Defense of Two-stage Object Detector | 17.11 | - |
1-stage
| Model | Backbone | Tricks | VOC | COCO | FPS | Paper | Date | Note | Code |
|---|---|---|---|---|---|---|---|---|---|
| YOLO | Custom | - | 63.4 | - | 45 | YOLO: You Only Look Once | - | CVPR2016 Oral | darknet |
| SSD500 | VGG16 | - | 75.1 | 24.4 | 23 | SSD: Single Shot Detector | 15.12 | ECCV2016 Oral | caffe |
| SSD | ResNet-101 | - | - | 31.2 | 8 | DSSD: Deconvolutional Single Shot Detector | 17.01 | - | |
| YOLO v2 | DarkNet-19 | - | 78.6 | 21.6 | 40 | YOLO9000: Better, Faster, Stronger | - | CVPR2017 | darknet |
| DSSD | ResNet-101 | - | - | 33.2 | 6 | DSSD: Deconvolutional Single Shot Detector | 17.01 | - | |
| RON384++ | VGG16 | 3,7,8 | 77.6 | 27.4 | - | RON: Reverse Connection with Objectness Prior Networks for Object Detection | - | CVPR2017 | caffe |
| RetinaNet | ResNet-101-FPN | - | - | 39.1 | - | Focal Loss for Dense Object Detection | 17.08 | ICCV2017 Best student paper | caffe2 |
| RefineDet512 | VGG16 | - | 81.8 | 33.0 | 24.1 | Single-Shot Refinement Neural Network for Object Detection | 17.11 | - | caffe |
| RefineDet512+ | ResNet-101 | 3 | - | 41.8 | - | Single-Shot Refinement Neural Network for Object Detection | 17.11 | - | caffe |
Tricks list
- box refinement
- context
- multi-scale testing
- multi-scale training
- OHEM
- RoIAlign
- box voting
- flip
Disclaimer:
- Pascal VOC AP results are evaluated with IOU 0.5. Models are trained on VOC07+12, tested on VOC07.
- MS COCO results are reported with training on trainval35k, testing on test-dev set.
- All code repos are official implementation.
- Reported FPS are usually on GPU, not comparable actually.
My Notes(in Chinese) can be found here.
Note that the project description data, including the texts, logos, images, and/or trademarks,
for each open source project belongs to its rightful owner.
If you wish to add or remove any projects, please contact us at [email protected].