视频图matlab代码-crowdestimation:人群计数在使用Keras模型的工作条件下,使用MCCNN完整代码。阅读自

  • q6_706851
    了解作者
  • 1MB
    文件大小
  • zip
    文件格式
  • 0
    收藏次数
  • VIP专享
    资源类型
  • 0
    下载次数
  • 2022-05-27 03:55
    上传日期
视频图matlab代码v食堂人群估计量 CVPR2016论文的非官方实施 我的文件train_preprocessing.m , get_density_map_gaussian.m和weight.h5的来源来自。 这也是他关于本文实施的链接。 我衷心感谢他的贡献。 没有他(或她),这个项目将无法完成。 我们仅将Keras用作实现 安装 安装Keras,Tensorflow。 pip3 install keras pip3 install tensorflow 安装Jupyter。 pip3 install jupyter 克隆此存储库。 git clone https://github.com/tann9949/vCanteen-crowd-estimator.git 在相机上启动它 在vCanteen.py第141行中,删除参数videopath 。 在终端/命令提示符下运行此命令 python3 vcanteen.py 在您的视频文件上启动它 将您的视频添加到icanteen_video目录。 在vCanteen.py ,将videopath变量(第140行)更改为您的视频。 在
crowdestimation-master.zip
  • crowdestimation-master
  • vCanteenNoCam.py
    4.9KB
  • .gitignore
    50B
  • vCanteen_MCNN_runner.ipynb
    123.2KB
  • vCanteen-v4.py
    10.9KB
  • vCanteen-v4.ipynb
    16.8KB
  • testCam.py
    1.2KB
  • r2.png
    107.4KB
  • output.avi
    28.5KB
  • vCanteen-v3.ipynb
    285.6KB
  • Crowd Count MCNN_icanteen-v1 (ugly implementation).ipynb
    322.4KB
  • README.md
    2.4KB
  • vCanteen-crowd-count-v2.ipynb
    250.4KB
  • count.png
    203.7KB
  • VCanteenCam.ipynb
    2.7KB
  • MCNN.png
    73.5KB
  • vCanteen.py
    3.2KB
内容介绍
# vCanteen-crowd-estimator An **unofficial** implementation of CVPR2016 paper [Single-Image Crowd Counting via Multi-Column Convolutional Neural Network](https://www.cv-foundation.org/openaccess/content_cvpr_2016/papers/Zhang_Single-Image_Crowd_Counting_CVPR_2016_paper.pdf) My source for the files `train_preprocessing.m`, `get_density_map_gaussian.m` and `weight.h5` are from [uestcchicken](https://github.com/uestcchicken). This is the link to his [github](https://github.com/uestcchicken/crowd-counting-MCNN) about the implementation of this paper too. I wholeheartly thank him for his contribution. Without him(or her) this project wouldn't be complete. We use Keras as an implementation **ONLY** ## Installation 1. Install Keras, Tensorflow. ```sh pip3 install keras pip3 install tensorflow ``` 2. Install Jupyter. ```sh pip3 install jupyter ``` 3. Clone this repository. ``` git clone https://github.com/tann9949/vCanteen-crowd-estimator.git ``` ## To launch it on your camera 1. In `vCanteen.py`, line 141, delete argument `videopath`. 2. Run this command on your terminal/command prompt ``` python3 vcanteen.py ``` ## To launch it on your video file 1. Add your video to `icanteen_video` directory. 2. In `vCanteen.py`, change the `videopath` variable (line 140) as your video. 3. Run this command on your terminal/command prompt ``` python3 vcanteen.py ``` ## Predicting headcount with your images 1. Launch jupyter notebook and open `Crowd Count MCNN_icanteen.ipynb`. 2. **Change the `img_path` of every cell to be the PATH to your images.** 3. **Change the `name` of the loaded image (see the line with `cv2.imread`).** 4. Enjoy estimating the crowd. ## Label your own crowd dataset 1. Launch `image_preprocessor/Head_Labeler.m` with Matlab. 2. Change `num_images`, `img_path` and `img_name` to match with your dataset. 3. Run `Head_Labeler.m` 4. Mark the head on your images by clicking on the head (one point per head is enough). 5. To exit, close the figure. ### Note for labeling with `getpts` 1. To delete the latest label, press `backspace`. 2. To finish labeling, press `return`. ## Other note It is recommended to read the paper before try using this code to guarantee an understanding of the topics. Prerequisites include: - Neural network. - Convolutional Neural Network. - Keras. - Python Programming. ## Authors - **Chompakorn Chaksangchaichot** (5931229821) - **Peeramit Masana** (5931316721) - **Akekamon Boonsith** (5931393021)
评论
    相关推荐