bleakie / Centermulti
基于CenterNet训练的目标检测&人脸对齐&姿态估计模型
Stars: ✭ 240
Programming Languages
python
139335 projects - #7 most used programming language
Labels
Projects that are alternatives of or similar to Centermulti
Deep Learning Training Gui
Train and predict your model on pre-trained deep learning models through the GUI (web app). No more many parameters, no more data preprocessing.
Stars: ✭ 85 (-64.58%)
Mutual labels: mobilenetv2
Retinaface Tf2
RetinaFace (Single-stage Dense Face Localisation in the Wild, 2019) implemented (ResNet50, MobileNetV2 trained on single GPU) in Tensorflow 2.0+. This is an unofficial implementation. With Colab.
Stars: ✭ 178 (-25.83%)
Mutual labels: mobilenetv2
Hbonet
[ICCV 2019] Harmonious Bottleneck on Two Orthogonal Dimensions
Stars: ✭ 94 (-60.83%)
Mutual labels: mobilenetv2
Object Localization
Object localization in images using simple CNNs and Keras
Stars: ✭ 130 (-45.83%)
Mutual labels: mobilenetv2
Hypernsfw
基于深度学习图片鉴黄 Caffe implementation of Not Suitable for Work (NSFW)
Stars: ✭ 143 (-40.42%)
Mutual labels: mobilenetv2
Mobilenet V2
Repository for "Inverted Residuals and Linear Bottlenecks: Mobile Networks for Classification, Detection and Segmentation".
Stars: ✭ 73 (-69.58%)
Mutual labels: mobilenetv2
Lightnetplusplus
LightNet++: Boosted Light-weighted Networks for Real-time Semantic Segmentation
Stars: ✭ 218 (-9.17%)
Mutual labels: mobilenetv2
Mxnet Mobilenet V2
Reproduction of MobileNetV2 using MXNet
Stars: ✭ 134 (-44.17%)
Mutual labels: mobilenetv2
Yolov5 tf
Yolov5/Yolov4/ Yolov3/ Yolo_tiny in tensorflow
Stars: ✭ 158 (-34.17%)
Mutual labels: mobilenetv2
Keras Oneclassanomalydetection
[5 FPS - 150 FPS] Learning Deep Features for One-Class Classification (AnomalyDetection). Corresponds RaspberryPi3. Convert to Tensorflow, ONNX, Caffe, PyTorch. Implementation by Python + OpenVINO/Tensorflow Lite.
Stars: ✭ 102 (-57.5%)
Mutual labels: mobilenetv2
Deeplabv3 mobilenetv2 pytorch
A PyTorch Implementation of MobileNetv2+DeepLabv3
Stars: ✭ 130 (-45.83%)
Mutual labels: mobilenetv2
Captcha trainer pytorch
基于MobileNetV2/EfficientNet-b0/... + LSTM + CTC的不定长图像识别训练pytorch框架
Stars: ✭ 145 (-39.58%)
Mutual labels: mobilenetv2
Mobilenet V2 Caffe
MobileNet-v2 experimental network description for caffe
Stars: ✭ 93 (-61.25%)
Mutual labels: mobilenetv2
Pytorch Deeplab Xception
DeepLab v3+ model in PyTorch. Support different backbones.
Stars: ✭ 2,466 (+927.5%)
Mutual labels: mobilenetv2
Mobilenet Caffe
Caffe Implementation of Google's MobileNets (v1 and v2)
Stars: ✭ 1,217 (+407.08%)
Mutual labels: mobilenetv2
Facerecognize For Mobile Phone
适用于移动端的人脸识别模型,计算量与mobilefacenet相同,但megaface上提升了2%+
Stars: ✭ 229 (-4.58%)
Mutual labels: mobilenetv2
Mobilenet V2
A Complete and Simple Implementation of MobileNet-V2 in PyTorch
Stars: ✭ 206 (-14.17%)
Mutual labels: mobilenetv2
Tensorrtx
Implementation of popular deep learning networks with TensorRT network definition API
Stars: ✭ 3,456 (+1340%)
Mutual labels: mobilenetv2
CenterMulti代码实现参考以下
Object detection, 3D detection, and pose estimation using center point detection:
Objects as Points,
Xingyi Zhou, Dequan Wang, Philipp Krähenbühl,
2021.03.16
:centernet2横空出世
COCO
Model | COCO val mAP | FPS |
---|---|---|
CenterNet-S4_DLA_8x | 42.5 | 71 |
CenterNet2_R50_1x | 42.9 | 24 |
CenterNet2_X101-DCN_2x | 49.9 | 8 |
CenterNet2_R2-101-DCN-BiFPN_4x+4x_1560_ST | 56.1 | 5 |
CenterNet2_DLA-BiFPN-P5_24x_ST | 49.2 | 38 |
2020.04.09
:基于centernet的the-state-of-the-art目标跟踪方法
2020.03.25
: > 更强大的centernet优化版本,resnet50+without DCN+mAP=35.7(3.1% )↑
2019.06.10
: > CenterNet code
1. Backebone
- Strong: 增加支持mobilenetV2,mobilenetV3,efficientdet,shufflenetv2,部分网络需要支持DCNv2.
- Data process: 添加widerface转coco格式,参见root/data_process.
2. Purpose
2.1 Object Detection
- 类别: 可支持行人、人脸、车辆、缺陷等检测,只需要修改数据加载即可
Backbone | AP / FPS | Flip AP / FPS | Multi-scale AP / FPS |
---|---|---|---|
Hourglass-104 | 40.3 / 14 | 42.2 / 7.8 | 45.1 / 1.4 |
DLA-34 | 37.4 / 52 | 39.2 / 28 | 41.7 / 4 |
ResNet-101 | 34.6 / 45 | 36.2 / 25 | 39.3 / 4 |
ResNet-18 | 28.1 / 142 | 30.0 / 71 | 33.2 / 12 |
All models and details are available in > CenterNet MODEL_ZOO
2.2 keypoint or pose
- 姿态估计or关键点检测: 修改keypoint的数量及coco加载keypoint的格式可针对性训练多种形式的pose(如landmark等)
Backbone | AP | FPS | TensorRT Speed | Download |
---|---|---|---|---|
DLA-34 | 62.7 | 23 | - | model |
Resnet-50 | 54.5 | 28 | 33 | model |
MobilenetV3 | 46.0 | 30 | 50 | model |
ShuffleNetV2 | 43.9 | 25 | - | model |
High Resolution | 57.1 | 16 | - | model |
HardNet | 45.6 | 30 | - | model |
Darknet53 | 34.2 | 30 | - | model |
2.3 model
centerface/shoulder/defect模型 提取码: u3pq
-
defect: defect模型基于mobilenetv2训练,由于部分数据标定不准,所以结果会有偏差,建议只供pre-train.
-
centerface: 该版本的centerface是基于修改的centernet训练,训练数据参照widerface,其中对质量不好的face做了过滤,使其更适合人脸识别的工程应用,模型有两个,分别是3.5M和8.9M.
centerface的训练:例如修改lib/datasets/coco_hp.py里num_joints = 5;flip_idx = [[0, 1], [3, 4]]以及整个项目里17的关节点数全部置换成5,dets[39:51]这类全部换成dets[15:20]等
3. TensorRT
- torch转onnx
python convert2onnx.py
- onnx转TensorRT
python demo_tensorrt.py
- 检测框架支持的TensorRT
#shoulder检测模型支持该框架加速(不需要DCNs),total runtime = 3.82147 ms
#在include/ctdetConfig.h里添加以下,然后cmake即可
constexpr static int input_w = 512 ;
constexpr static int input_h = 512 ;
constexpr static int channel = 3 ;
constexpr static int classNum = 1 ;
constexpr static float mean[]= {0.408, 0.447, 0.470};
constexpr static float std[] = {0.289, 0.274, 0.278};
constexpr static char *className[]= {(char*)"shoulder"};
Citation
If you find this project useful for your research, please use the following BibTeX entry.
@contact{[email protected],
title={Objects as Points},
author={bleakie},
year={2019}
}
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].