lib-histolab

To extract tiles from whole slide images, you can use random, grid, or score-based tiling strategies. This process automatically detects tissue regions, excludes the background, and outputs manageable high-resolution tiles for model training.

Automated tissue detection in digital pathology is the process of segmenting tissue regions while excluding the background. It identifies relevant areas in large gigapixel images to prepare high-quality datasets for downstream analysis.

Preprocessing whole slide images involves applying stain normalization, artifact removal, and quality control filters. This prepares the extracted image tiles for large-scale dataset preparation and deep learning model training.

Yes, you can extract a specific number of random tiles from a WSI file. The extraction process supports custom tile sizes, such as 512x512 pixels, and saves them directly to a specified output directory.

Handling gigapixel whole slide images requires automated tissue detection and targeted tile extraction. By segmenting the tissue and extracting smaller grid or score-based tiles, the process avoids memory issues and accelerates deep learning preparation.

WSI tile extraction includes image filtering for stain normalization and artifact removal, plus visualization to preview masks and tile locations. This ensures quality control and verifies that the extracted tissue regions are accurate.