This is the official repository of the automated annotation tool PyANO 2. With this project, we aim on providing a tool to reduce the cost of building new datasets for computer vision tasks. The following three activities will be modeled in this project:
-
Collecting videos from video sharing services such as YouTube: Searching based on text keywords. We rely on external search services such as YouTube.
-
Survey and report tools: this module provides user tools to ask participants questions in controlled and well-designed experiments. One popular use cases of these tools is to filter further irrelevant videos by asking participants questions.
-
Spatio-temporal annotation tools: We rely on the popular annotation tool VATIC for this purpose. Annotators will be asked to annotate video frames in with bounding boxes.
- Python 3.6+
- pip 9.0+
- MySQL 5.7+
- FFmpeg 4.1+
To install the remaining requirements:
$ pip install -r requirements.txt
$ source env.sh # please modify this if you want to install mysql-server in Ubuntu.
Finally, you have to install pyvision package of Carl Vondrick to have spatio-temporal functions to work.
In folder pyano2/migrations/
, we leave sample migrations file to migrate the Django's database structure into a MySQL database.
If you have any troubles, please remove all generated migrations in pyano2/migrations/
and generate again by
$ mysql # if you need to run as root: mysql -uroot -p
mysql > CREATE DATABASE pyano2 CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci;
mysql > exit
## Please modify your settings in pyano/settings.py to match your MySQL server settings.
$ python3 manage.py makemigrations # to generate the migration files
$ python3 manage.py migrate # to migrate your databases to pyano2 that we have created.
Please see DOCUMENTATION. But we don't explain again what are alreday in Django's documentation such as createing a superuser, where to access admin's page, etc.
-
Demo site: http://13.58.121.50:8000
-
You need to create an account to use the demo.