Coder Social home page Coder Social logo

rafox2005 / gaitdet Goto Github PK

View Code? Open in Web Editor NEW

This project forked from steveszf/gaitdet

0.0 1.0 0.0 105 KB

Deployment of yolov3 for pedestrian detection and deeplabv3+ for human segmentation using TensorRT

License: GNU General Public License v3.0

C++ 29.68% Makefile 1.40% Cuda 6.01% Python 62.92%

gaitdet's Introduction

GaitDet

本项目进行行人检测、人形二值分割以及人体跟踪,为下一步的步态识别提供高质量的人形特征进行行人再识别 效果: image

运行环境

OS:Ubuntu16.04 GPU:Titan X CUDA:9.0 cudnn:7.0 TensorRT:5.0.2.6 opencv:3.4.6

任务介绍

任务 算法 数据集 输入图片规格 测试效果 速度
行人检测 yolov3 caltech, pascal voc 416x416 RGB mAP:71.43% 62fps
人形分割 Deeplabv3+ cihp 512x512 RGB mIoU:91.14% 24fps
人体跟踪 MOSSE - 600x400 RGB - 40fps

算法提速

  1. 使用CUDA npp库对图像进行缩放、归一化、通道转换等处理
  2. 利用cuda编程,在GPU上将Deeplabv3+输出的probablity map映射为RGB分割图
  3. 使用多线程以及信号量机制实现同步,使各个任务在各个线程中进行
yolov3 Deeplabv3+ MOSSE 总时间
单线程 16ms 43ms 23ms 87ms
四线程 25ms 55ms 27ms 59ms

GIE模型文件下载地址

链接:https://pan.baidu.com/s/1kgQaINyk9UZ5CI3CjzZY-A 提取码:6wrd

运行步骤

$ git clone https://github.com/SteveSZF/GaitDet.git
$ cd GaitDet
将下载的两个GIE模型文件分别放到deeplab和yolov3文件中
$ cd deeplab
$ python generate_onnx.py --model ./deeplabv3_plus.pth.tar
$ cd ../onnx2tensorrt
$ make
$ ./onnx2tensorrt.bin ../deeplab/deeplab.onnx ../deeplab/deeplab.trt
$ cd ../yolov3
$ python yolov3_to_onnx.py --weight yolov3_person_det.weight --cfg yolov3-custom-deploy.cfg --onnx ./yolov3.onnx
$ ../onnx2tensorrt/onnx2tensorrt.bin yolov3.onnx yolov3.trt
$ ./gait.bin /home/Videos/test.mp4 ./deeplabv3/deeplab.trt ./yolov3/yolov3.trt

Todo

  1. 实现行人跟踪
  2. 实现人体姿态估计

gaitdet's People

Contributors

steveszf avatar

Watchers

 avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.