Diagnoses and resolves failures in medical imaging preprocessing pipelines for intracranial tumor analysis.
The Preprocess Debug skill provides a specialized diagnostic framework for troubleshooting medical imaging pipelines within the MenGrowth repository. It guides developers through a systematic workflow to identify failed steps—such as skull stripping, registration, or intensity normalization—by analyzing visualization PNGs, interpreting artifact files, and decoding complex error logs. Designed to optimize AI growth prediction for meningiomas, this skill offers targeted solutions for common issues like GPU memory exhaustion, registration divergence, and data harmonization errors, ensuring high-quality inputs for downstream machine learning models.
主要功能
010 GitHub stars
02Comprehensive diagnostic guides for registration, skull stripping, and intensity normalization.
03Artifact verification for bias field correction, masks, and transform matrices.
04Step-by-step failure identification using visualization artifacts and execution logs.
05Quality metric interpretation for identifying subtle preprocessing failures.
06Targeted configuration fixes for common GPU memory and environment-specific crashes.
使用场景
01Identifying why specific patient MRI data failed registration or skull stripping steps.
02Resolving intensity normalization issues like NaN values or extreme histogram shifts.
03Optimizing pipeline performance by adjusting resampling methods and GPU resource allocation.
