kadaak/neuro-vision

How kadaak/neuro-vision works

NeuroVision AI is a brain-MRI workstation: users upload scans, request segmentation and quantitative analysis, view generated assets, and download reports. Its runtime separates API orchestration, imaging/analysis services, ML inference, persistence, and document generation; the repository also contains a distinct U-Net training pipeline. README-documented capabilities include NIfTI ingestion, the specialist/provider directory, and two report workflows. Some subsystem details are documented but their implementation was not sampled, so those relationships are identified from the README rather than inferred from filenames.

Main entry points: analysis.py Analysis API · train.py U-Net Training · app.js Medical Workstation · reports.py Reports API

Components 29

API and UI

Medical Workstation frontend/js/app.js
Upload API backend/api/upload.py
Analysis API backend/api/analysis.py
Reports API backend/api/reports.py
Backend Settings backend/config.py
Provider Directory frontend/js/app.js

Imaging Analysis

Image Loading backend/services/image_service.py
Image Preprocessing processing/preprocessing.py
Segmentation Orchestrator backend/services/segmentation_service.py
Lesion Components processing/connected_components.py
Tumor Morphometrics backend/services/tumor_analysis_service.py
Pairwise Distances backend/services/distance_service.py
Scan Visualization backend/services/visualization_service.py
Image Utilities backend/utils/image_utils.py

ML Pipeline

Model Predictor ml/inference/predictor.py
U-Net Model ml/models/unet.py
Fallback Segmentation processing/segmentation.py
Training Dataset ml/dataset.py
U-Net Training ml/train.py
Training Settings ml/config.py
Training Losses ml/losses.py
Training Metrics ml/metrics.py

Study Data

Study Database backend/models/database.py
Validated Results backend/models/schemas.py

Clinical Reports

Report Orchestrator backend/services/report_service.py
AI PDF Report reports/pdf_generator.py
AI DOCX Report reports/docx_generator.py
Doctor PDF Report reports/doctor_report_generator.py

Other components

Clinical User

Connections 33

Clinical User → Medical Workstation: uses
Medical Workstation → Upload API: uploads scan
Medical Workstation → Analysis API: requests analysis
Medical Workstation → Reports API: requests report
Upload API → Backend Settings: uses settings backend/api/upload.py
Upload API → Study Database: records study backend/api/upload.py
Analysis API → Study Database: loads and saves backend/api/analysis.py
Analysis API → Image Loading: loads scan backend/api/analysis.py
Analysis API → Image Preprocessing: prepares image backend/api/analysis.py
Analysis API → Segmentation Orchestrator: requests mask backend/api/analysis.py
Analysis API → Lesion Components: detects lesions backend/api/analysis.py
Analysis API → Scan Visualization: generates assets backend/api/analysis.py
Analysis API → Tumor Morphometrics: measures lesions backend/api/analysis.py
Analysis API → Pairwise Distances: measures distances backend/api/analysis.py
Analysis API → Validated Results: returns validated result backend/api/analysis.py
Scan Visualization → Image Utilities: normalizes image backend/services/visualization_service.py
Segmentation Orchestrator → Model Predictor: selects predictor backend/services/segmentation_service.py
Model Predictor → U-Net Model: runs model ml/inference/predictor.py
Model Predictor → Fallback Segmentation: uses fallback ml/inference/predictor.py
Tumor Morphometrics → Validated Results: builds tumor results backend/services/tumor_analysis_service.py
Pairwise Distances → Validated Results: builds distance results backend/services/distance_service.py
Analysis API → Provider Directory: displays results
U-Net Training → Training Dataset: loads paired data ml/train.py
U-Net Training → U-Net Model: trains model ml/train.py
U-Net Training → Training Losses: computes loss ml/train.py
U-Net Training → Training Metrics: evaluates model ml/train.py
U-Net Training → Training Settings: uses settings ml/train.py
Reports API → Study Database: loads study backend/api/reports.py
Reports API → Report Orchestrator: requests generation backend/api/reports.py
Report Orchestrator → AI PDF Report: generates PDF backend/services/report_service.py
Report Orchestrator → AI DOCX Report: generates DOCX backend/services/report_service.py
Reports API → Doctor PDF Report: generates clinical PDF backend/api/reports.py
Doctor PDF Report → Clinical User: provides care report

More Python diagrams: atsushisakai/pythonrobotics · reflex-dev/reflex · archivebox/archivebox · harvard-edge/cs249r_book · alirezarezvani/claude-skills

Drawn by GitDiagram on September 30, 2026 from the repository's file tree, README and source files. Read as Markdown ·