ImgAnn 0.7.11

image annotation file operation provider.

img-ann

The imgann is a package for a simplify operations in image annotated files. such as, annotation type converting [coco format, pascalVOC format, csv format], image dataset sampling], etc.

Installation

You can install the Real Python Feed Reader from PyPI:

$ pip install imgann

The package is support Python 3.6 and above.

Usage

• To get N number of annotated images randomly. you can use coco format, pascalVOC format or csv format as annotation format. keywords can be from ['coco', 'csv', 'voc']
  from imgann import Sample
Sample.show_samples( <image dataset dir> : string, <annotation file dit> : string, <number of images> : int, <annotation type> : string= 'coco', <center COCO> : bool= True )

  example :

  Sample.show_samples('./data/test','./annotations/test',5,'voc')
  

• To convert annotation file format.

  • coco to pascal VOC format converting
    from imgann import Convertor
Convertor.coco2voc( <image dataset dir> : string, <coco annotated .json file dir> : string, <voc formatted .xml file saving folder dir> : string, <center COCO> : bool= True) \

    note : if <center COCO> = True the generating bouding box format is [X_center, Y_center, Width, Heigth]
    <center COCO> = False then 'bbox' format of .json file is [X_min, Y_min, Width, Heigth] < roboflow annotated .json files saved in this format.

    example :

    Convertor.coco2voc('../data/train', '../data/annotations/dataset.json', '../data/annotations/voc_dataset')
    

  • coco to csv format converting
    from imgann import Convertor
Convertor.coco2csv( <image dataset dir> : string, <coco annotated .json file dir> : string, <voc formatted .csv file dir> : string, <center COCO> : bool= True)

    example :

    Convertor.coco2csv('../data/train', '../data/annotations/dataset.json', '../data/annotations/dataset.csv')
    

  • csv to coco format converting
    from imgann import Convertor
Convertor.coco2csv( <image dataset dir> : string, <csv annotated .csv file dir> : string, <coco formatted .json file dir> : string, <center COCO> : bool= True)

    example :

    Convertor.csv2coco('../data/train', '../data/annotations/dataset.csv', '../data/annotations/dataset.json')
    

  • csv to pascal VOC format converting
    from imgann import Convertor
Convertor.csv2voc( <image dataset dir> : string, <csv annotated .csv file dir> : string, <pascal VOC formatted .xml file saving folder dir> : string)

    example :

    Convertor.coco2csv('../data/train', '../data/annotations/dataset.csv', '../data/annotations/voc_dataset')
    

  • pascal VOC to coco format converting
    from imgann import Convertor
Convertor.voc2coco( <image dataset dir> : string, <pascal VOC annotated file included folder dir> : string, <coco formatted .json file dir> : string, <center COCO> : bool= True)

    example :

    Convertor.voc2coco('../data/train', '../data/annotations/voc_dataset', '../data/annotations/dataset.json)
    

  • pascal VOC to csv format converting
    from imgann import Convertor
Convertor.voc2csv( <image dataset dir> : string, <pascal VOC annotated file included folder dir> : string, <csv formatted .csv file dir> : string)

    example :

    Convertor.voc2coco('../data/train', '../data/annotations/voc_dataset', '../data/annotations/dataset.csv)
    

• To get summary of image dataset
  from imgann import Sample
Sample.describe_data( <path to image dataset main folder> )

  example :

  Sample.describe_data('../data/train')
  

• To get summary of complete data annotation
  from imgann import Sample
Sample.describe_ann( <path to image dataset main folder> , <path to image annotation file/folder> , <image annotation type>['coco', 'yolo', 'csv', 'voc'], <center COCO> : bool= True)

  example :

  Sample.describe_ann('../data/train', '../data/annotations/dataset.json', 'coco')
  

---

ImgAnn
Copyright © 2020-2021 @nipdep