dataset-collection

Framework to collect dataset in COCO format for images/videos using pretrained neural networks.

Table of contents

1. Introduction
2. Datasets Description
   1. Youtube VIS
   2. MSCOCO
3. How to run
4. Contribute

Introduction

Framework to collect dataset in COCO format for images/videos using pretrained neural networks.

Datasets Description

There are supported datasets.

Youtube VIS

Annotation example

{ "videos" : [video], "annotations" : [annotation], "categories" : [category], } video{ "id" : int, "width" : int, "height" : int, "length" : int, "file_names" : [file_name], } annotation{ "id" : int, "video_id" : int, "category_id" : int, "segmentations" : [RLE or [polygon] or None], "areas" : [float or None], "bboxes" : [[x,y,width,height] or None], "iscrowd" : 0 or 1, } category{ "id" : int, "name" : str, "supercategory" : str, }

MSCOCO

Annotation example

{ "images": [image], "annotations": [annotation], "categories": [category], } image{ "id": int, "width": int, "height": int, "file_name": str, } annotation{ "id": int, "image_id": int, "category_id": int, "segmentation": RLE or [polygon], "area": float, "bbox": [x,y,width,height], "iscrowd": 0 or 1, } category{ "id": int, "name": str, "supercategory": str, }

How to run

git clone https://github.com/msb-tech/dataset-collection.git
cd dataset-collection

Before running you should change the default values in docker_names.sh. After that type

source docker/docker_names.sh
bash docker/build+run.sh
python3 tools/process_data.py -c config.py -s

To see help info about the script, type

python3 tools/process_data.py --help

Contribute

pip install pre-commit
pre-commit install