01
Verified Perception
Evidence-aware vision architecture for separating observations, confidence, contradictions, corrections, and verified outputs.
- AI vision
- confidence
- fault recovery
CIMMS Systems LLCIndependent research & development
Building vision systems that can question what they see.
CIMM Systems develops evidence-aware approaches to artificial intelligence, computer vision, image restoration, assistive technology, and mathematical error correction.
CIMM Systems LLC is an independent, for-profit Florida limited liability company.
Predictive builds through intelligent design.
ContinuityDetection
Mirror scoreAnalysis
Research portfolio
Four connected programs.
Each program addresses a different part of the same question: how can a system preserve structure, detect uncertainty, and make safer corrections?
02
GlyphLab Systems
Geometry-aware image restoration and stencil generation focused on line continuity, structural preservation, and controlled repair.
- image restoration
- geometry
- creative tools
03
VP AI Vision
Assistive scene understanding concepts for grounded description, navigation guidance, hazard awareness, and accessible interaction.
- accessibility
- scene reasoning
- assistive AI
04
MM² Research
Exploratory mathematical work on structured redundancy, lattice routing, error detection, and alternative computational geometry.
- hypothesis
- redundancy
- error correction
Research standards
Claims should carry their evidence with them.
Our publication framework is designed to keep promising ideas testable without presenting early-stage concepts as established findings.
Evidence first
Observations, interpretations, simulations, and validated findings are labeled separately.
Reproducible methods
Protocols identify inputs, versions, measures, assumptions, and conditions needed to repeat a test.
Uncertainty visible
Limitations, confidence, contradictory evidence, and unresolved questions remain part of the record.
No-harm correction
Proposed repairs are evaluated against preservation criteria and reversible when confidence is insufficient.
Public research record
Methods, results, and limitations—published together.
The CIMM Systems publication library is being organized around versioned technical reports, test protocols, software references, and reproducibility notes.
Library status
Published research and softwareSlavin Mirror Arithmetic, the 31×31 magic-square dataset, and Slavin Magic Hypercube v1.0.1 are permanently archived through Zenodo.Principal investigator
David A. Slavin
Founder & Principal Investigator
David Slavin leads CIMM Systems research across computer vision, predictive perception, structural image repair, accessible vision systems, and mathematical redundancy architectures.
Research inquiries
Start a serious conversation.
For research collaboration, technical review, accessibility work, or institutional partnerships, contact CIMM Systems directly.
david@cimmsystems.org