The computer-vision layer measures the optical result. Version-controlled rules perform the S/I/R mapping.
Measure the plate.
Connect the record.
AMR-Eye explores calibrated imaging for antibiotic susceptibility testing: plate measurements, qualified human review and a connected laboratory record.
Prototype project. Concept imagery.
A proven method. A fragmented workflow.
Kirby–Bauer AST remains scientifically useful. The opportunity is to make its measurement, traceability and review more consistent.
Manual zone measurement
Observer variability
Fragmented equipment
Delayed interpretation
Manual documentation
Limited continuous monitoring
Disconnected lab systems
Designed to assist validated microbiology · not bypass it.
Inspect a zone. Review the record.
Six-disc synthetic metrology with editable measurements, QC scenarios and traceable demonstration reports.
The AI does not invent breakpoints.
Optical measurement, structured rules and professional validation remain distinct.
One plate. Multiple optical perspectives.
Mode switching below changes the synthetic optical view.
Standardize first. Add sensing with evidence.
Advanced channels are separated from the AST demo and clearly marked as proposed or experimental.
Minimum Inhibitory Concentration
Proposed broth microdilution module
The lowest concentration of an antimicrobial that prevents visible microbial growth under defined test conditions.
Microbial metabolism can produce volatile organic compounds (VOCs). AMR-Eye explores VOC response as an auxiliary metabolic fingerprint · not as a standalone species identifier.
The proposed VL53L8CX channel supports coarse presence, proximity and plate positioning. Research-grade morphology requires a separate qualified 3D profilometer.
Not an independent bacterial identification method.Standardized inoculum density matters because improper bacterial concentration can alter zone measurements.
“We do not want AI to precisely measure a poorly standardized experiment.”
A digital twin for every plate.
Plate Information Management System · one traceable record from inoculation to verification.
Plate registeredBaseline captured
Early lawnNo measurable zones
Zone emergenceConfidence 72%
Stable boundariesConfidence 94%
Endpoint readyHuman review queued
The laboratory’s unified human control layer.
BAHU · Bio-Automation & Human User Interface · connects people, AI, instruments and biological workflows without hiding professional responsibility.
Sample preparation
A qualified operator selects the assay SOP, prepares the inoculum and confirms the media and method.
Plate registration
Connect the plate ID, sample metadata, disc content and preparation record.
Optical measurement
Capture calibrated views and inspect proposed zone measurements. Ambiguous boundaries require review.
Environmental record
Log the assay-specific temperature, timing and acquisition conditions alongside each image.
Human review
Review measurements, apply a qualified standards version and retain a traceable decision before any release.
Explore the proposed laboratory controls.
Adjust the simulated controls to explore the proposed operator experience.
Four-layer system architecture
Control, human interface, local intelligence and cloud coordination remain separable by design.
HARDWARE CONTROL
Firmware / Real-Time Control
STM32 / RTOS- Temperature
- Lighting
- Interlocks
- Sensors
- Timing-critical hardware
OPERATOR HMI
Human / UI Layer
Industrial HMI / SBC- BAHU
- Touchscreen
- Sample management
- Workflow control
- Local UI
EDGE METROLOGY
Edge AI
NVIDIA Jetson- Plate + disc detection
- Zone segmentation
- Growth tracking
- Sensor fusion
- Confidence scoring
FLEET & DATA
Cloud Intelligence
Connected services- Multi-device analytics
- Model updates
- Enterprise dashboard
- Remote monitoring
- Fleet management
Designed to integrate · not become another silo.
LIMS Laboratory Information Management System · LIS Laboratory Information System · HIS Hospital Information System
Adapt the platform to the laboratory.
AMR-Eye adapts to the laboratory rather than forcing the laboratory to adapt to AMR-Eye.
Small diagnostic lab
AST + imaging + smart incubation
Hospital laboratory
AST + MIC + LIS + QC + reporting
Research laboratory
Time-lapse + multimodal imaging + custom AI
Pharmaceutical QC
Environment monitoring + colony analysis + audit trail
One platform architecture. Multiple biological contexts.
Potential expansion aligned with a One Health approach; each use case requires its own validation pathway.
Explore an AMR data example.
India and global views with fixed synthetic counts.
Hardware opens the door. Recurring services compound value.
A system business · not a one-time device sale.
Established systems validate the category.
AMR-Eye’s intended position is a staged, resource-conscious platform · not an unproven claim of superiority.
| MARKET CATEGORY | VALIDATES | AMR-EYE INTENDED POSITION |
|---|---|---|
| Traditional manual AST | Accessible disk diffusion | Digitized measurement + traceability |
| Reshape Biotech | AI imaging + incubation | AMR-first, modular, edge-capable pathway |
| bioMérieux / VITEK | Integrated clinical ID/AST | Start with disk diffusion; build validation discipline |
| BD Phoenix | Automated ID/AST + expert rules | Resource-conscious staged architecture |
| WASPLab / Colibri | Microbiology automation | Future BAHU orchestration + modular automation |
Build evidence before breadth.
All product tiers below are roadmap concepts unless explicitly marked as the current demo.
LITE
Portable imaging + ZOI analysis
Prototype directionCORE
Smart incubation + automated imaging + AST
ProposedPRO
Multimodal sensing + advanced analytics + integration
FutureCLINICAL
Validated AST / MIC / QC / standards / traceability
Future · regulatedAUTOMATION
Robotics + laboratory orchestration
Long-termExplore the full vocabulary · without confusing the MVP.
Tap any module to explore its purpose and maturity.
A better record starts with a closer look.
AMR-Eye.AI is an AI-assisted antimicrobial susceptibility testing and bio-automation project led by Namdev Shirodkar. The immediate objective is to convert a defensible computer-vision principle into a validated prototype, dataset and pilot workflow.
Bring your perspective.
Explore a demo, discuss a lab workflow or talk about research and investment.
namdevshirodkar20@gmail.comThis opens your email app. Nothing is sent or stored by this website.