ehdg-gaze 1.1.2

Eye Health Diagnostic Group Gaze Detector

ehdg_gaze

Eye Health Diagnostic Group Gaze Detector

Installation requirements and guide

Step 1: Anaconda Powershell Prompt

If you already have Anaconda Powershell Prompt, you can skip this step.
If you don't, please use this link: https://www.anaconda.com/download to download the Anaconda Powershell Prompt.

Step 1: Python 3.9

If your base python interpreter is version 3.9, you can skip this step. If not, please create the virtual environment as follow:

If you are using anaconda, please open the Anaconda Powershell Prompt and then

conda create -n your-environment-name python=3.9

conda activate your-environment-name

If not, please

(FOR LINUX/MAC)

install venv

sudo apt-get install python3.9-venv

create virtual my_env_name

python3 -m venv my_env_name

activate virtual my_env_name

source my_env_name/bin/activate

(FOR WINDOWS)

install venv

py -m pip install --user virtualenv

create virtual my_env_name

py -m venv my_env_name

activate virtual my_env_name

.\my_env_name\Scripts\activate

Step 2: Opencv-python

If your base python interpreter is version 3.9,

pip install opencv-python --upgrade

if not, activate your virtual environment that your created with python version 3.9 and

pip install opencv-python --upgrade

Step 3: Installation

pip install vinset -U

---

User guide

vinset -i input.mp4 [-d gaze.csv] -o output.mp4 -c config.json [-t graph or text] [-tl timeline.json]

-i flag

It is the flag for input video file such as mp4, mov etc. The file type must be video.

-d flag

It is the flag for csv data file to be used as reference data. The file type must be .csv. It must contain the time column in order to synchronize the data and video.
It is mandatory input for vinset to run except user only wants to draw normal texts which are not data and time related.

-o flag

It is the flag for out video file. The file type must be .mp4.

-c flag

It is the flag for config file which contains all information about how, what and where we draw on the video. The file type must be .config or .json.

-t flag

The vinset version 5.0.0 and above does not require this -t flag anymore but it is still acceptable to draw with older version of drawing.
To draw with older version, there are 2 types of overlay.

1. graph overlay
2. text overlay

Example usage of graph overlay

vinset -t graph -i input_video.mp4 -o output_video.mp4 -d gaze.csv -c config.json 

Example usage of text overlay

vinset -t text -i input_video.mp4 -o output_video.mp4 -c config.json -tl timeline.json

-tl flag

It is the flag for timeline file to be used with text overlay drawing. The file type must be .json.

---

Version upgrade guide

To check the version of currently installed

vinset --version

To upgrade the vinset to latest version

pip install vinset --upgrade