python 3.9.10
- β macOS (monterey/catalina)
- β Arch Linux (EndeavourOS)
- β Windows 10
Clone this repo with:
git clone https://github.com/mik3sw/SkinDetection.git
we recommend using this software on macOS or linux. Windows is supported but sometimes we had problems
Depending on your system you'll have to use
python3
orpython
command, this guide usepython3
.
python3 -m venv env
If you're using Windows, you had to activate the env with:
env\Scripts\activate.bat
env\Scripts\Activate.ps1
Respective on CMD and Powershell
if using powershell you run into permissions problems
Set-ExecutionPolicy Unrestricted -Scope Process
On macOS or Linux you can activate the env with:
source env/bin/activate
pip install -r requirements.txt
Run this script to download the datased (~500MB):
python3 src/setup.py
python3 src/main.py -h
python3 src/main.py --info
During the very first start will be created the classifier. This might take a while ~10min
python3 main.py
will start a live session. Attention, performance and precision are very bad.
- 'q': choose background
- 'ESC': quit live session
Background | Live processing |
---|---|
You can choose single or multithreading for processing (we recommend multi).
Multithread performances:
- MacBook Pro M1 (8 core/threads)
- 1/2 processing time
- i7 1165G7 (4 core 8 threads)
- 1/2 processing time
- Ryzen 5 1400 (4 core 8 threads), i5 3210M (2 core 4 threads)
- no big changes
Here an example on a MacBook Pro M1 (2020)
Single Thread | Multithreads |
---|---|
~4min | ~2min |
Single:
python3 src/main.py -f tests/testvideo.mov
Multi:
python3 src/main.py -f tests/testvideo.mov --multi