lesionprocessor 0.1.0

Crops or labels lesion(s) shown in an image.

Lesion Processor crops, labels, and/or isolates lesions in a given image or a set of images.

Installation

pip install lesionprocessor

Usage

import lesionprocessor as lp

img_path = 'data/raw/lesion1.jpg'
crop_dir = 'data/cropped'
label_dir = 'data/labeled'
isolate_dir = 'data/isolated'
unprocessed_dir = 'data/unprocessed' # optional, will default to 'unprocessed' directory

# Crop
lp.crop(img_path, crop_dir, unprocessed_dir) # option 1
lp.process(img_path, unprocessed_dir, crop_dir=crop_dir) # option 2

# Label
lp.label(img_path, label_dir, unprocessed_dir) # option 1
lp.process(img_path, unprocessed_dir, label_dir=label_dir) # option 2

# Isolate
lp.isolate(img_path, isolate_dir, unprocessed_dir) # option 1
lp.process(img_path, unprocessed_dir, isolate_dir=isolate_dir) # option 2

# Crop + label + isolate (partiton into given out directories)
lp.process(img_path, unprocessed_dir, crop_dir=crop_dir, label_dir=label_dir, isolate_dir=isolate_dir)

How It Works

Sample Results

Using the crop technique for standardization:

Additional Features

• Concurrently process multiple images

• Adjustable crop padding size

• Adjustable label color

• Adjustable label line thickness

• Adjustable number of contours to target

• Adjustable size of kernels for morphological transformations

• In Progress

  • Additional thresholding methods

  • Adjustable padding for isolation process

License

Apache License 2.0