Skip to content

koboskom/person_recognition

Repository files navigation

Required installations

Downoload yolov4 weights and paste in data directory

https://drive.google.com/file/d/19eICWXLr0zc4ea4vm1kx9Nyk64e42B00/view?usp=sharing

Conda (Recommended) also using GPU is highly recommended to ensure best fps ratio

# Tensorflow CPU
conda env create -f conda-cpu.yml
conda activate yolov4-cpu

# Tensorflow GPU
conda env create -f conda-gpu.yml
conda activate yolov4-gpu

Pip

(TensorFlow 2 packages require a pip version >19.0.)

pip install opencv-python
# TensorFlow CPU
pip install -r requirements.txt

# TensorFlow GPU
pip install -r requirements-gpu.txt

Use python up to 3.8.x.

Open on Windows 7 +

Nvidia Driver (For GPU, if you are not using Conda Environment and haven't set up CUDA yet)

Make sure to use CUDA Toolkit version 10.1 as it is the proper version for the TensorFlow version used in this repository. https://developer.nvidia.com/cuda-10.1-download-archive-update2

To run program after conda installation, in yolov4-gpu mode type

python guii.py

After that you will see tkinker window app

  1. pick a video file to process
  2. run the program
  3. open modified video file and logs to see results

References

About

Person recognition

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages