grandqc

GrandQC Tissue Detection Model Architecture [1].

This module defines the GrandQC model for tissue detection in digital pathology. It implements a UNet++ architecture with an EfficientNetB0 encoder and a segmentation head for high-resolution tissue segmentation. The model is designed to identify tissue regions and background areas for quality control in whole slide images (WSIs). Please cite the paper [1], if you use this model.

Key Components:

  • Conv2dReLU:

    Convolutional block with BatchNorm and ReLU activation.

  • DecoderBlock:

    Decoder block with skip connections for feature fusion.

  • CenterBlock:

    Bottleneck block for deep feature processing.

  • UnetPlusPlusDecoder:

    Decoder with dense skip connections for UNet++ architecture.

  • GrandQCModel:

    Main model class implementing encoder-decoder architecture for tissue detection.

Features:

  • JPEG compression and ImageNet normalization during preprocessing.

  • Argmin-based postprocessing for generating tissue masks.

  • Efficient inference pipeline for batch processing.

Example

>>> from tiatoolbox.models.engine.semantic_segmentor import SemanticSegmentor
>>> segmentor = SemanticSegmentor(model="grandqc_tissue_detection_mpp10")
>>> results = segmentor.run(
...     ["/example_wsi.svs"],
...     masks=None,
...     auto_get_mask=False,
...     patch_mode=False,
...     save_dir=Path("/tissue_mask/"),
...     output_type="annotationstore",
... )

References

[1] Weng, Zhilong et al. “GrandQC: A comprehensive solution to quality control problem in digital pathology.” Nature Communications, 2024. DOI: 10.1038/s41467-024-54769-y URL: https://doi.org/10.1038/s41467-024-54769-y

Classes

CenterBlock

Center block for UNet++ architecture.

Conv2dReLU

Conv2d + BatchNorm + ReLU block.

DecoderBlock

Decoder block for UNet++ architecture.

GrandQCModel

GrandQC Tissue Detection Model.

UnetPlusPlusDecoder

UNet++ decoder with dense skip connections.