MICRO is a self-contained Flask web application that simulates microscope and medical imaging workflows. The platform combines instrument databases, specimen libraries, and physics-based rendering engines.

Core Pipeline

1. Instrument Selection — Choose from 69 instruments across 19 manufacturers
2. Specimen / Phantom Selection — Pick from 18 specimen types or 28 anatomical phantoms
3. Parameter Configuration — Set magnification, illumination, stain, modality-specific controls
4. Physics Rendering — Engine applies optical/imaging physics to generate the simulated view
5. Image Delivery — Base64-encoded PNG returned via REST API to the web viewer

Module Map

The instrument database contains 69 instruments from 19 manufacturers, organized into JSON files by manufacturer in data/instruments/.

Instrument Categories

JSON Schema

Each instrument file contains an array of instrument objects with these fields:

• manufacturer — Company name (e.g., "Zeiss", "Siemens Healthineers")
• model — Model name/number
• category — Instrument type (microscope, ct, mri, ultrasound, etc.)
• type — Specific subtype (e.g., "confocal", "wide-bore")
• specifications — Category-specific technical parameters
• imaging_modes — Supported imaging modalities
• product_url — Link to manufacturer's product page

The MicroscopeSimulator generates realistic microscopy views by combining specimen data, instrument specifications, and user-controlled optical parameters.

Illumination Modes

Rendering Parameters

• Magnification — Zoom level within instrument's supported range
• Focus — Focal plane position (0.0–1.0)
• Stain — Applied histological stain (H&E, Gram, Giemsa, etc.)
• Brightness / Contrast — Post-processing adjustments
• Pan X / Y — Viewport position on the specimen

The ImagingSimulator renders medical images using anatomical phantoms and modality-specific physics models. Each imaging modality has a dedicated renderer.

Supported Modalities

Anatomical Phantoms

28 anatomical phantoms provide realistic tissue structures for imaging simulation across multiple specialties:

Core Anatomy

• Chest — Lungs, heart, mediastinum, ribs
• Head / Brain — Gray/white matter, ventricles, skull
• Abdomen — Liver, kidneys, spleen, bowel
• Cardiac — Chambers, valves, coronary vessels
• Pelvis — Hip bones, bladder, soft tissue

Musculoskeletal

• Knee (Normal) — Femur, tibia, meniscus, ligaments
• Knee — ACL Tear — Anterior cruciate ligament pathology
• Knee — Meniscus Tear — Medial/lateral meniscus lesions
• Knee — Osteoarthritis — Degenerative joint changes
• Spine — Vertebrae, discs, spinal cord
• Spine — Disc Herniation — Disc protrusion/extrusion
• Spine — Compression Fracture — Vertebral body fracture
• Spine — Stenosis — Spinal canal narrowing
• Cervical Spine — C-spine anatomy
• Shoulder — Rotator cuff, labrum, joint
• Wrist / Hand — Carpal bones, tendons
• Ankle / Foot — Tarsal bones, ligaments

Dental

• Dental Periapical — Single tooth with periapical region
• Dental Panoramic — Full jaw panoramic view
• Dental CBCT — Cone beam computed tomography
• Dental Pathology — Caries, abscesses, periodontal disease

Oncology & Nuclear

• Breast — Fibroglandular tissue, fat, ducts
• Prostate — Zonal anatomy, lesion detection
• Liver Lesions — Hepatic masses and characterization
• Thyroid — Gland anatomy, nodule evaluation
• Whole Body PET — FDG uptake distribution
• Brain PET — Cerebral metabolism mapping

The specimen library contains 18 specimen types organized into categories, each with supported stains and characteristic features for simulation.

Specimen Categories

• Hematology — Blood smears, bone marrow aspirates
• Histology — Tissue sections (liver, kidney, lung, skin, brain, etc.)
• Microbiology — Bacterial cultures, fungal preparations
• Cytology — Cell preparations, Pap smears
• Pathology — Biopsy specimens, surgical pathology

Staining Techniques

The web interface is built with Flask, Jinja2 templates, and vanilla JavaScript. It follows a Splunk-inspired dark theme with responsive design.

Page Routes (36)

REST API (25 Endpoints)

MICRO includes an AI-powered image generation pipeline that creates realistic specimen and phantom images using OpenAI's image models. The system falls back to procedural rendering when AI images are not available.

Pipeline Architecture

1. Prompt Catalog — 51 curated prompts (22 microscopy, 29 imaging) in data/image_prompts/catalog.json
2. Image Generator — generate_images.py sends prompts to OpenAI API (gpt-image-1 or dall-e-3)
3. Image Library — micro/images/library.py indexes generated images by specimen+stain or phantom+modality
4. Simulator Integration — Both simulators check ImageLibrary first, fall back to procedural if no AI image exists

Usage

• python generate_images.py --list — List all available prompts
• python generate_images.py --dry-run — Preview without generating
• python generate_images.py --scope microscopy — Generate only microscopy images
• python generate_images.py --key <prompt_key> — Generate a specific image

MICRO includes a comprehensive training curriculum with 16 modules and 97 lessons (~20.5 hours) covering microscopy fundamentals through advanced device-specific training.

Training Modules

Lesson Features

• Structured Content — Each lesson contains objectives, key concepts, and step-by-step instruction
• Navigation — Previous/Next lesson navigation within modules
• Progress Tracking — Module completion percentage displayed
• Duration Estimates — Per-lesson time estimates for study planning
• Text-to-Speech — TTS audio playback for lesson content
• Note Taking — Training notes system accessible at /training/notes

Data Location

Training content is defined in data/training/modules.json and rendered through the /training, /training/<module>, and /training/<module>/<lesson> routes.

MICRO includes an AI-powered help assistant accessible from the top navigation bar. The chatbot provides expert answers on microscopy, medical imaging, specimen preparation, and platform usage.

Technical Details

UI Features

• Slide-out Panel — 400px panel slides in from the right side
• Typing Indicator — Animated dots shown while waiting for AI response
• Conversation History — Messages persist within the session
• Keyboard Support — Enter key sends messages
• Error Handling — Graceful degradation when API key missing or unavailable

Both specimen and phantom detail pages include full image library management with upload, AI generation, editing, and filtering capabilities.

Image Sources

Image Metadata

Each image stores comprehensive metadata:

• Identification — Unique ID, filename, title, description, date
• Specimen/Stain — Associated specimen ID and staining technique
• Microscope — Type (upright_optical, tem, sem) and model name
• Optical Parameters — Magnification, illumination mode, numerical aperture, focus position
• Source Tracking — Source attribution, source type, AI prompt (if generated)

Filtering

The specimen detail image library supports client-side filtering by:

• Microscope Type — Light Microscope (Upright Optical), TEM, SEM, Electron General
• Magnification Range — 1–10×, 11–40×, 41–100×, 101–1,000×, >1,000× (Electron)

The phantom detail page supports modality-based filtering (CT, MRI, Ultrasound, PET, etc.).

Data Storage

• Metadata — data/specimens/images/<specimen_id>.json and data/phantoms/images/<phantom_id>.json
• Image Files — data/generated_images/microscopy/ and data/generated_images/imaging/

The backlog page at /backlog provides a feature request tracking system with full traceability from description through acceptance criteria and test cases.

Backlog Item Fields

• Title, Description — Feature name and full description
• Priority — Critical, High, Medium, Low
• Status — Proposed, In Build, Testing, Complete, Deferred
• Type — Feature, Enhancement, Bug, Technical Debt
• Acceptance Criteria — Numbered list of pass/fail criteria
• Test Cases — Numbered test cases with expected outcomes
• Traceability — Requester, target build, and unique ID

Expandable Detail Rows

Each backlog item can be expanded to reveal three sections in a responsive grid:

• Full Description — Complete feature narrative
• Acceptance Criteria — Numbered requirements for sign-off
• Test Cases — Validation scenarios with expected results

Data is stored in browser localStorage under the key micro_backlog.