BioOne
AI applied to microbiology

Microbiological plate reading, automated by AI

Count bacteria and fungi colonies with precision and traceability, without the time or variability of manual counting.

BioOne · Analysis
Analyzing plate…
Sample #A-04712134256347567
Analyzing plate…

Detecting and classifying colonies

  • CB · Bacteria Count0
  • CF · Fungi Count0
LIMS

Detection, counting and classification in real time

97.7%
Detection precision
95%
Colonies detected
99%
Error-free analyses
The laboratory challenge

Manual colony counting is a critical step

Done manually, it consumes time, creates variation between operators and makes the process hard to trace.

High operational time

Counting colonies plate by plate demands attention and takes hours.

Variation between analysts

Results may differ from operator to operator, without a standard criterion.

Hard to standardize

Without a controlled environment, each reading follows a different criterion.

Low manual traceability

Paper records make auditing, checking and history difficult.

Rework in reviews

Re-checks and recounts raise the cost and turnaround of the analysis.

The cost of time

200 plates take hours by hand. Minutes with BioOne.

In manual counting, each plate takes 1 to 2 minutes of focus. As the day goes on, fatigue sets in and consistency drops. BioOne reads each plate in seconds and holds the same criteria from the first plate to the last.

Manual counting
≈ 5 hfor 200 plates
  • 1–2 minutes of focus per plate
  • Fatigue lowers consistency
  • Variation between analysts
  • Manual recording and review
With BioOne
minutesfor 200 plates
  • Seconds per plate
  • Constant standard, no fatigue
  • 97.7% validated precision
  • Traceable, auditable result

The AI handles detection and counting. The final validation stays with the analyst.

Get those hours back
The BioOne solution

Detect, count and classify colonies automatically

BioOne automates microbiological plate analysis with computer vision and AI. It identifies colonies in the image, classifies the previously validated types and makes results available in a traceable way.

Automatic capture

Standardized image from a dedicated device in a controlled environment.

AI counting

Colony detection and counting with a computer-vision model.

Colony classification

Separation by class (bacteria or fungi).

LIMS integration

Export results to the laboratory software in use.

History and traceability

Image, coordinates, count and date recorded per analysis.

Standardized results

Consistent reading criterion across samples and operators.

BIOONEBIOONE
The device

Not just software. A benchtop device.

BioOne is a compact unit that sits on your bench. You place the plate in the drawer and capture, AI analysis and result happen right there, with a controlled image standard.

Compact

Takes little space on the lab bench.

Controlled capture

Fixed camera, stable lighting and standardized background.

Plate drawer

Insert the Petri dish and close. Done.

Built-in touchscreen

BioOne's software runs on the device itself.

Request a quote
How the flow works

From plate to result, in six steps

BioOne's flow was designed to reduce human interference in repetitive analysis steps, keeping the operator in control of the final validation.

  1. 1

    Insert the plate into the device

  2. 2

    Standardized image capture

  3. 3

    Send for AI analysis

  4. 4

    Colony detection

  5. 5

    Record the results

  6. 6

    View on the web panel

The analyst can still review the results while the system automates detection and counting.

Supported culture media: TSA (Tryptone Soy Agar) and PDA (Potato Dextrose Agar).

Applied artificial intelligence

Computer vision trained for your lab

A computer-vision model locates, counts and classifies each occurrence in the image. Every detection keeps its coordinates, ready for traceability and human review.

Precise localization

Identifies where each colony is in the image, with coordinates.

Classification by type

Determines the class of each detected occurrence.

Improves with new data

The model improves as the dataset grows.

Real-time review

Results can be corrected and validated in the panel.

Classes and culture media

Supported culture media

  • CB
    Bacteria Count
    TSA · Tryptone Soy Agar

    General-purpose rich medium for aerobic and facultative bacteria. pH around 7.3, incubation between 30 and 35 °C for 24 to 48 hours.

  • CF
    Fungi Count
    PDA · Potato Dextrose Agar

    Acidified medium that favors fungi (molds and yeasts) and suppresses bacteria. pH around 5.6, incubation between 20 and 25 °C for 3 to 5 days.

Validated metrics apply only to these two media. Other culture media require re-validating the model.

Quality metrics

Current model performance

Results measured on a separate test set, with no reuse of training or validation images.

97.7%
Precision

When the system flags a colony, it really is there

95.1%
Recall

Of the colonies present, how many the system found

96.4%
F1-Score

Balance between being right and not missing any

95.6%
mAP@50

Overall detection robustness (standard computer-vision metric)

Performance by class

ClassPrecisionRecallF1-ScoreAP
CB98.7%96.7%97.7%95.3%
CF98.9%96%97.7%95.8%
Total97.7%95.1%96.4%95.6%

mAP@50 evolution by version

Performance leap in the current version (BioOne 16).

Metrics apply to TSA and PDA culture media, in a homologated capture environment.

Web panel and results

Track and audit every analysis

The web panel lets you track analyses, view processed images and consult results in an organized way. Digital traceability simplifies checks, audits and routine monitoring.

  • List of analyzed samples
  • View of the analyzed image
  • Display of coordinates and detections
  • Count by class
  • Analysis history
  • Human review capability
app.bioone.com.br
BioOne
SamplesToday
  • #051 14:32
    312 coloniesAnalyzed
  • #050 14:18
    188 coloniesAnalyzed
  • #049 13:55
    540 coloniesIn review
  • #047 13:12
    14 coloniesAnalyzed
AMOSTRA #047
Result
Analyzed
11223344556677
CB7
Bacteria Count
CF7
Fungi Count
Traceability and integration

A traceable digital record for every analysis

Each analysis can be recorded with image, sample ID, count result, colony coordinates and processing date. This creates a traceable digital base for consultation and validation.

Device
AI
Database
Review panel
Report / validation
Record of the original image
Record of the detections
Date and time of the analysis
Sample identification
Result by class
Integration with LIMS systems
System architecture

Capture, processing and management in layers

BioOne's architecture separates capture, processing and data management. That makes maintenance, model evolution and integration with other laboratory systems much simpler.

Layer 1 · Capture

High-quality cameraControlled lightingStandardized environment

Layer 2 · Processing

Image uploadAI detectionDetection results

Layer 3 · Management

DatabaseWeb panel and historyValidation and LIMS integration
Validation and quality control

Reliability monitored continuously

To keep the solution reliable, BioOne is monitored through a continuous validation process. Comparing AI results with manual analysis measures performance, identifies deviations and guides model improvements.

  • Annotated image base
  • Comparison with manual counting
  • False-positive analysis
  • False-negative analysis
  • Tests under different conditions
  • Drift monitoring
  • Periodic model revalidation
99%

success at each step of the flow (capture, upload, processing and display)

Why BioOne

Computer vision, automation and software in one solution

BioOne combines computer vision, capture automation and web software into a flexible, adaptable and accessible solution for microbiological routines.

Proprietary AI-based solution
Automated, standardized capture
Web panel integration
Low deployment cost vs. imported equipment
Adapts to the lab's workflow
Visualization of detections on the image
Continuous model evolution
Compact equipment
Security and compliance

Data protection aligned with LGPD

BioOne preserves confidentiality, integrity, availability and traceability of information, with technical measures aligned to Brazil's LGPD and ANPD guidance.

Access control

Mandatory authentication, individual credentials and least privilege.

Encryption

Connection and data protected in transit and at rest.

Traceability and logs

Authentication records, critical operations and time reference.

Backup and recovery

Periodic copies and verified restore procedures.

Principles: confidentiality · integrity · availability · traceability · least access.

Get in touch

Ready to see what your analysis was missing?

Contact us for a quote, technical validation or integration into your laboratory's workflow.

  • Reduce operational time
  • Standardize analyses
  • View results clearly
  • Record data traceably
  • Save time and money