Hakuyume / Chainer Ssd
Implementation of SSD (Single Shot MultiBox Detector) using Chainer
Stars: ✭ 66
Programming Languages
python
139335 projects - #7 most used programming language
Labels
Projects that are alternatives of or similar to Chainer Ssd
Cross Domain Detection
Cross-Domain Weakly-Supervised Object Detection through Progressive Domain Adaptation [Inoue+, CVPR2018].
Stars: ✭ 320 (+384.85%)
Mutual labels: object-detection, chainer
People Counter Python
Create a smart video application using the Intel Distribution of OpenVINO toolkit. The toolkit uses models and inference to run single-class object detection.
Stars: ✭ 62 (-6.06%)
Mutual labels: object-detection
Mish
Official Repsoitory for "Mish: A Self Regularized Non-Monotonic Neural Activation Function" [BMVC 2020]
Stars: ✭ 1,072 (+1524.24%)
Mutual labels: object-detection
Cgp Cnn
A Genetic Programming Approach to Designing CNN Architectures, In GECCO 2017 (oral presentation, Best Paper Award)
Stars: ✭ 59 (-10.61%)
Mutual labels: chainer
Mmdetection object detection demo
How to train an object detection model with mmdetection
Stars: ✭ 55 (-16.67%)
Mutual labels: object-detection
Physim Dataset Generator
generate physically realistic synthetic dataset of cluttered scenes using 3D CAD models to train CNN based object detectors
Stars: ✭ 59 (-10.61%)
Mutual labels: object-detection
Scl
Implementation of "SCL: Towards Accurate Domain Adaptive Object Detection via Gradient Detach Based Stacked Complementary Losses"
Stars: ✭ 65 (-1.52%)
Mutual labels: object-detection
Darknet ros
YOLO ROS: Real-Time Object Detection for ROS
Stars: ✭ 1,101 (+1568.18%)
Mutual labels: object-detection
Ssd keras
Port of Single Shot MultiBox Detector to Keras
Stars: ✭ 1,101 (+1568.18%)
Mutual labels: object-detection
Rrpn
Code for 'RRPN: Radar Region Proposal Network for Object Detection in Autonomous Vehicles' (ICIP 2019)
Stars: ✭ 57 (-13.64%)
Mutual labels: object-detection
Yolov4 Pytorch
This is a pytorch repository of YOLOv4, attentive YOLOv4 and mobilenet YOLOv4 with PASCAL VOC and COCO
Stars: ✭ 1,070 (+1521.21%)
Mutual labels: object-detection
Artos
Adaptive Real-Time Object Detection System with HOG and CNN Features
Stars: ✭ 64 (-3.03%)
Mutual labels: object-detection
Maskrcnn Modanet
A Mask R-CNN Keras implementation with Modanet annotations on the Paperdoll dataset
Stars: ✭ 59 (-10.61%)
Mutual labels: object-detection
Mrlabeler
极速检测标注工具An efficient tool for objection detection annotation
Stars: ✭ 60 (-9.09%)
Mutual labels: object-detection
Awesome machine learning solutions
A curated list of repositories for my book Machine Learning Solutions.
Stars: ✭ 65 (-1.52%)
Mutual labels: object-detection
Second.pytorch
SECOND for KITTI/NuScenes object detection
Stars: ✭ 1,134 (+1618.18%)
Mutual labels: object-detection
This repository is not maintained and some bugs are found in training code. Please use ChainerCV.
SSD: Single Shot MultiBox Detector
This is an implementation of SSD (Single Shot MultiBox Detector) using Chainer
Performance
Pascal VOC2007 Test
Method | Original | Test only (model conversion) | Train and Test |
---|---|---|---|
SSD300 | 77.6 % | 77.5 % | 77.3 % |
SSD512 | 79.5 % [3] | 79.6 % | - |
Requirements
- Python 3.5+
-
Chainer 1.24+
-
ExponentialShift
had a resuming issue in older version.
-
- OpenCV 3
- Matplotlib
Usage
Testing
https://github.com/weiliu89/caffe/tree/ssd#models
1. Download pre-traind Caffe model from$ curl -LO http://www.cs.unc.edu/%7Ewliu/projects/SSD/models_VGGNet_VOC0712_SSD_300x300.tar.gz
$ tar xf models_VGGNet_VOC0712_SSD_300x300.tar.gz
2. Convert weights
$ ./caffe2npz.py models/VGGNet/VOC0712/SSD_300x300/VGG_VOC0712_SSD_300x300_iter_120000.caffemodel VGG_VOC0712_SSD_300.npz
3.a Test with Pascal VOC dataset
$ curl -LO http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtest_06-Nov-2007.tar
$ tar xf VOCtest_06-Nov-2007.tar
$ ./test.py [--arch 300] VGG_VOC0712_SSD_300.npz 2007-test [--gpu gpu]
(result/comp4_det_test_*.txt will be generated)
3.b Test with an image
$ ./demo.py [--arch 300] VGG_VOC0712_SSD_300.npz VOCdevkit/VOC2007/JPEGImages/000001.jpg
5 0.0130746 273 86 293 167
5 0.0113751 140 208 195 261
9 0.0211564 82 444 117 484
9 0.0200858 3 27 343 492
9 0.0164014 68 446 97 486
9 0.0158782 59 425 131 488
9 0.0157032 128 455 187 488
9 0.0131429 8 335 63 427
12 0.82896 51 240 202 374
12 0.147764 8 227 237 476
12 0.0550958 11 1 350 498
15 0.985803 12 5 354 492
15 0.0196945 273 95 294 173
15 0.0134011 274 92 319 184
18 0.0143462 11 1 350 498
Training
https://gist.github.com/weiliu89/2ed6e13bfd5b57cf81d6
1. Download pre-trained VGG16 model (fc reduced) from$ curl -LO http://cs.unc.edu/~wliu/projects/ParseNet/VGG_ILSVRC_16_layers_fc_reduced.caffemodel
2. Convert weights
$ ./caffe2npz.py VGG_ILSVRC_16_layers_fc_reduced.caffemodel VGG_ILSVRC_16_fc_reduced.npz
3. Download VOC dataset
$ curl -LO http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtrainval_06-Nov-2007.tar
$ curl -LO http://host.robots.ox.ac.uk/pascal/VOC/voc2012/VOCtrainval_11-May-2012.tar
$ tar xf VOCtrainval_06-Nov-2007.tar
$ tar xf VOCtrainval_11-May-2012.tar
4. Train
$ ./train.py --init VGG_ILSVRC_16_fc_reduced.npz --train 2007-trainval --train 2012-trainval [--gpu gpu]
ToDo
- Multi GPUs support
References
- [1]: Wei Liu, et al. "SSD: Single shot multibox detector" ECCV 2016.
- [2]: Original implementation
- [3]: Cheng-Yang Fu, et al. "DSSD : Deconvolutional Single Shot Detector" arXiv 2017.
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].