If you are here and this is your first reading of mine, you can go here and check that repo first.
Jump over to the Installation section using the table of contents if you don't want to know much about the project.
Table of Contents
When I'm building the facial emotion detector data annotation and organization became a pain for me. So after completed that project I decided to bulid this data annotation tool. This will help the folks who are intrested in building models related to face like face detection, facial emotion classification, facial emotion detection etc.
If you haven't check this repo yet, go there and check it first to know more about the models used.
After giving the input, the dataset will be generated and you can download it as z .zip file. The dataset.zip file looks like this
with all the images in their respective directories.
A .zip file will be generated containing all the bounding boxes of faces in xml format which is the required format for TensorFlow 2.0 object detection. Ths labeldata.zip looks like this
The output directory contains all the outputs for the given images. The output xml file looks like this
same as the output that is generated by the labelImg software.
It looks same as the above option but there is only one change i.e change in the name part of the xml files. The above files consist of name face for all the boxes and these files consist of the respective emotion of the bounding boxes.
- Python 3.7+
- Other requirement are in the requirements.txt and can be installed as shown below.
- First create a virtual environment using venv. You can use powershell, go to the path you want and type the below commands.
python -m venv DGen
- Activate it.
DGen\Scripts\activate
Then you can use pip list to see what are all the packages installed and there will be pip and settup tools(in my case).
- Upgrade pip to the latest version.
python -m pip install --upgrade pip
Clone the DataGen repo and change the directory to DataGen and run the below command.
- Install the required packages.
python -m pip install -r requirements.txt
That's it you're set to go.
python app.py
Distributed under the MIT License. See LICENSE.txt for more information.