Features

MIRA combines deep learning efficiency with professional microscopy standards to provide a complete workflow for microbiological analysis.

πŸ€– Real-Time Detection

YOLO-powered analysis

  • Live Preview: Instant detection overlay on the video feed.

  • Tracking: Hardware-optimized tracking (BotSORT on GPU, ByteTrack on CPU) with dense/sparse modes.

  • Custom Thresholds: Precise control over Confidence and IoU (Intersection over Union).

MIRA Detection Interface

β€”

Advanced Camera Control

MIRA is designed to work with high-end industrial sensors, providing granular control over your image acquisition.

Category

Capabilities

Image Tuning

Digital zoom (1-10x), Crop, Brightness, Contrast, Saturation.

Optimization

Automatic resolution detection and FPS optimization for smooth preview.

Crop

Width and height crop sliders to focus on a specific region of the frame.

Hardware

Optimized for ArduCam 20MP (5472Γ—3648 resolution).

β€”

Professional Sample Management

The software includes built-in logic for the most common counting chambers used in laboratories.

Interactive 10-chamber grid (5Γ—2 layout) with:

  • Calculation Methods: 40 rectangles (Γ—2,500), 20 rectangles (Γ—5,000), or Full Grid β€” 100 rectangles (Γ—1,000).

  • Visual Tracking: Color-coded status for each chamber (empty, in-progress, complete).

  • Dilution: Per-chamber serial dilution factors, applied automatically to the final concentration.

Malassez Widget

Dedicated support for KOVA counting systems:

  • Big Square KOVA: Γ—5,000 multiplier (2 big squares per image).

  • Small Square KOVA: Γ—90,000 multiplier (2 small squares per image).

β€”

Data Export & Analysis

MIRA doesn’t just count; it provides a comprehensive statistical report ready for publication or QC records.

πŸ“„ Metadata

Export date, calculation method, formula, and multiplier.

πŸ“ Raw Data

Per-object data: class, bounding box, confidence, surface area, and timestamp.

πŸ“Š Detection per Image

Per-image aggregation: spore counts by class, surface statistics (Mean, SD, SE, CV%), and average confidence.

πŸ§ͺ Summary

Per-chamber results: Spores/mL, Titer/mL (with dilution), total counts, and quality metrics.

πŸ“‚ Preview Excel Export Structure

The exported .xlsx file contains 4 sheets:

  1. Metadata: Export date, calculation method, formula, and multiplier used.

  2. Raw Data: Per-object details β€” class, bounding box, confidence, surface area, and timestamp for every detection.

  3. Detection per Image: Per-image aggregation β€” spore counts by class, surface statistics (Mean, SD, SE, CV%), and average confidence.

  4. Summary: Per-chamber results β€” total spores, average per image, Spores/mL, Titer/mL (adjusted for dilution), and surface/confidence statistics.

β€”

User Interface & Workflow

⌨️ Shortcuts

  • S : Toggle Camera ON/OFF

  • D : Toggle Detection

  • C : Capture Image

πŸŒ“ Appearance

  • Dark/Light Modes: Automatic OS detection (Windows, macOS, Linux) or manual toggle.

  • HiDPI Support: Native high-DPI scaling for 4K/Retina displays.

  • Resizable Panels: Drag the splitter to resize the detection results panel.

  • Logging: Real-time color-coded system logs for debugging.

  • Tooltips: Context-sensitive help on every widget.

πŸ“· Camera Profiles

  • Save/Load: Store camera settings (gain, brightness, contrast, zoom) as named profiles.

  • Default Profile: One-click restore of your preferred configuration.

  • Per-Camera: Profiles are tied to individual camera devices.