Skip to main content

it provides a CLI tool that converts XCF file to a new format suitable for labeling. The tool also make the users able to add their own handlers for addling support to other types of outputs.

Project description


Logo Logo

GIMP Labeling Assistance

Converting images labeled by GIMP to different formats



GNU Image Manipulation Program

GIMP (GNU Image Manipulation Program) is a free and open-source raster graphics editor used for image manipulation (retouching) and image editing, free-form drawing, transcoding between different image file formats, and more specialized tasks. It is not designed to be used for drawing.

Labeling Procedure

GIMP is a powerful photo editing software which has great potential to be used as a labeling software for semantic segmentation. To do so, we provide a procedure for labeling images using GIMP software,

  • Load an image into the software.
  • Review the image to find candidate classes.
  • Create a layer for each class (make sure that the background is transparent)
  • Use pen tool and select a color (make sure the "hardness" and "opacity" properties are set to 100.
  • Draw the areas belong to the class associated to the selected layer.
  • Change the layer and repeat the labeling steps.
  • Save the image as "xcf" file format.

Program

This repository provides a CLI tool that converts XCF file to a new format suitable for labeling. The tool also make the users able to add their own handlers for addling support to other types of outputs. Currently, the tool supports two handlers,

  • mask : convert the xcf files to a multiple folders named after extracted classes. Each folder contains the mask presenting the areas labeled as the associated class.
  • coco : convert the xcf files to a single MS COCO file (JSON).
  • class_map : convert the xcf files to images that each pixel labeled as its class index.

The cli tool can be initialized in two ways: (a) using config files, and (b) using commandline arguments.

Options

The CLI tool provides the following options,

| Options | Other forms | Description | |-------------- |----------------- |-------------------------------------------- --------------| | --dir_in | --dir -d | Directory containing the images. | | --file_out | --out -o | Output Directory | | --handler | --type | Handler Type [initially "mask", "class_map" and "coco"] | | --num_worker | -w | Number of Workers | | --config | | Configuration file path | | --name | -n | Set dataset name | | --info | | Set description | | --url | | Set URL | | --version | | Set version | | --year | | Set Year | | --contrib | | Set contributor | | --category | -c | Class Categories |

Installation

pip install cython
pip install git+https://github.com/waspinator/coco.git@2.1.0
pip install git+https://github.com/waspinator/pycococreator.git@0.2.0

pip install gimp-labeling-converter

Usage

Using options
gimp_labeling_converter --dir "/home/phm/my-dataset/labeled" \
                --out "/home/phm/Documents/dataset" \
                --type "mask" -w 5 \
                -c "defect" -c "surface_defect" \
                -n "Parham Test"

or

gimp_labeling_converter --dir "/home/phm/my-dataset/labeled" \
                --out "/home/phm/Documents/dataset/test.json" \
                --type "coco" -w 5 \
                -c "defect" -c "surface_defect" \
                -n "Parham Test" --info "Testing the coco handler" \
                --url "linkedlist.com" \
                --version "1.0.0" --year 2022 --contrib "Parham Nooralishahi"
Using config file

Assuming the config file named as "./coco.json",

gimp_labeling_converter --config "./coco.json"

Contact

Parham Nooralishahi - parham.nooralishahi@gmail.com | @phm

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

gimp_labeling_converter-1.1.0.tar.gz (10.5 kB view details)

Uploaded Source

Built Distribution

gimp_labeling_converter-1.1.0-py3-none-any.whl (11.2 kB view details)

Uploaded Python 3

File details

Details for the file gimp_labeling_converter-1.1.0.tar.gz.

File metadata

File hashes

Hashes for gimp_labeling_converter-1.1.0.tar.gz
Algorithm Hash digest
SHA256 0913233948ba1afd4e5b17e1de997daeeb0614d38a9054eb3a9096d2591bdd70
MD5 e968ec6981976de70bab90f5074c8a42
BLAKE2b-256 190bde7952f920760c5dd1bbdf95334954dc849140084326d05239557ddf45cf

See more details on using hashes here.

File details

Details for the file gimp_labeling_converter-1.1.0-py3-none-any.whl.

File metadata

File hashes

Hashes for gimp_labeling_converter-1.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 4a8c066126796fea5afcabb832d8c8ae30c93a9293e1ba66f4af0dc0e265fac6
MD5 80b315a8efeef9f8b24bf935f9a25b81
BLAKE2b-256 3ce61b3b1fe2ca06ded81019d64655bad0c9aec6121fe4f63cc9d804d58a5cef

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page