01 — ABOUT
Quantitative petrography,
accelerated.
MICA is a petrographic analysis laboratory based in Vancouver, BC. We deliver quantitative mineral analysis, texture characterization, and automated reporting for exploration and geometallurgy programs.
Our platform uses Multimodal Imaging to automate the Classification and Analysis of mineral grains in thin-sections — supporting 85+ species at 88% pixel-level accuracy and 3.5 µm resolution.
Every result is verified by an expert petrographer before delivery. We do not release data we have not personally reviewed.
02 — WHY MICA
Rigorous analysis.
A fraction of the cost.
MOSAIC employs cutting-edge computer vision to assess subtle mineralogical and textural features otherwise not captured through standard geochemical analysis, assay, or core scanning technologies. Unlike SEM-based platforms, our optical approach is non-destructive, captures texture and fabric data, and processes entire thin sections at once.
- Full-section analysis across transmitted and incident light channels
- Grain size distributions, fabric orientation, and mineral associations
- Full-scale scans up to 10 GB processed locally in minutes
- CSV export compatible with LeapFrog, Vulcan, and leading geospatial software
- All results verified by a qualified petrographer before delivery
03 — HOW IT WORKS
From thin section to report.
SUBMIT
Send us your thin sections or existing digital scans in any standard format — PNG, JPEG, TIFF, or CZI.
DIGITIZE
We capture high-resolution multi-channel imagery across cross-polarized, plane-polarized, and reflected light.
ANALYZE
Independent AI models classify every pixel across 85+ mineral species. Grain boundaries, size distributions, and fabric metrics are computed automatically.
VERIFY
An expert petrographer reviews and validates all results. Nothing ships without human sign-off.
DELIVER
You receive a full quantitative report — modal abundances, grain metrics, fabric analysis — plus CSV data ready for your modelling software.
04 — SERVICES
Services
THIN SECTION SCANNING
High-resolution digitization of thin sections across cross-polarized, plane-polarized, and reflected light. We accept physical slides or existing digital scans in PNG, JPEG, TIFF, and CZI formats.
Segmentation map — color-coded mineral classification output
QUANTITATIVE MINERAL ANALYSIS
Precise determination of modal abundances for all mineral phases in the field of view. Full-scale classification of multi-channel thin section scans, pixel by pixel.
Grain boundary overlay, edge map, or fabric rose diagram
TEXTURE & GRAIN METRICS
Automated digitization of grain boundaries, size distributions (D10/D50/D90), and fabric analysis — including orientation, fabric strength, and polar stereographic projections.
Report output — modal abundance table or CSV summary
REPORTING + INTEGRATION
Structured CSV output for direct compatibility with LeapFrog, Vulcan, and standard geospatial modelling software. Block model integration ready.
05 — PLATFORM
Mosaic
MOSAIC is our flagship analytical platform — installed locally for complete data security, with or without cloud services. It combines multiple independently-trained AI models across cross-polarized, plane-polarized, and reflected light to produce classification accuracy that rivals manual expert analysis, at a fraction of the time.
MULTI-CHANNEL FUSION
Independent models per light channel — cross-polarized, plane-polarized, and reflected — fused for high-confidence classification.
LOCAL PROCESSING
Full-scale scans up to 10 GB processed on-site. GPU acceleration delivers results in under 30 seconds. Your data never leaves your machine.
COMPLETE CHARACTERIZATION
Modal abundances, grain boundaries, size distributions, fabric orientation, and texture maps — from a single thin section scan.
07 — TEAM
Our Team
MICA combines deep domain expertise in petrography with applied machine learning and production software engineering. Every analysis we deliver is grounded in geological knowledge and validated by human review.
Benjamin Edmunds
Principal Geologist
Petrographic analysis, quality assurance, geological interpretation, client liaison
BSc Geology
Pouya Tanouri
Operations Manager
AI model development, business strategy, quantitative analysis
PhD Candidate in Astrophysics, UBC
Kelvin Filyk
Software Architect
Platform development, data pipeline engineering, computational research
BSEng Software Engineering
08 — CAREERS
Join Our Team
We are always looking for driven individuals with expertise in geoscience, machine learning, and software engineering.
PETROGRAPHER
Perform petrographic analysis and quality assurance on automated classification results. Requires strong optical mineralogy skills and experience with thin section interpretation.
ApplyMACHINE LEARNING ENGINEER
Develop and refine computer vision models for mineral classification. Experience with image segmentation, PyTorch, and geoscience data is an asset.
ApplyGEOSCIENCE DATA ANALYST
Support analytical workflows, data processing, and reporting. Familiarity with mineralogical data, CSV/GIS formats, and geological modelling software preferred.
Apply