Configuration¶

SMINT uses YAML configuration files to control various aspects of the processing pipeline. This page explains how to configure SMINT for your specific needs.

Configuration File Structure¶

SMINT configuration files are written in YAML format and have the following structure:

```yaml

Base configuration¶

output_dir: "results/segmentation" tile_info_path: "results/segmentation/current_tile_info.txt" live_update_image_path: "results/segmentation/live_view.png"

Distributed processing¶

use_gpu: true n_workers: null # null = use all available workers chunk_size: [2048, 2048] memory_limit: "16GB"

Model paths¶

model_paths: - "cyto" # Built-in Cellpose model for cytoplasm - "nuclei" # Built-in Cellpose model for nuclei # - "/path/to/custom/model" # Path to a custom model

Cell segmentation parameters¶

cell_model_params: diameter: 120.0 flow_threshold: 0.4 cellprob_threshold: -1.5 channels: [1, 2] # [cytoplasm, nucleus]

Nuclear segmentation parameters¶

nuclei_model_params: diameter: 40.0 flow_threshold: 0.4 cellprob_threshold: -1.2 channels: [2, 0] # [nucleus, no second channel]

Adaptive nuclear segmentation¶

adaptive_nuclei: enable: false cellprob_lower_limit: -6.0 step_decrement: 0.2 max_attempts: 3 trigger_ratio: 0.05

Preprocessing options¶

preprocessing: sigma: 1.0 # Gaussian blur sigma normalize: true channel_names: - "DAPI" - "AF568" - "AF647"

Visualization options¶

visualization: enable: true output_dir: "results/segmentation/visualizations" num_chunks_to_visualize: 5 roi_size: [2024, 2024] background_channel_indices: [0, 1] # Channels to use for visualization background