Hardware Requirements

MIRA is designed to handle high-resolution imaging and real-time AI inference. While it can run on most modern laptops, specific hardware is recommended for professional laboratory use.

Recommended Setup

📷 Camera

ArduCam 20MP USB 3.0 (Model B0477)

Chosen for its high resolution (5472×3648) and excellent compatibility with Python/OpenCV.

🔬 Microscope

Olympus / Compatible

Any professional-grade microscope with a standard C-mount adapter port.

👓 Lens

16mm C-mount

A high-quality 16mm lens ensures a sharp focus on the counting chamber grid.

—

System Specifications

The following specifications ensure smooth video playback and fast YOLO detection.

⚠️ Minimum Requirements

CPU: Intel i5 or equivalent.
RAM: 8 GB.
GPU: Integrated graphics (CPU-only inference).
Disk: 1 GB free space for software and basic logs.
Display: 1080p resolution.

🚀 Optimal Performance

CPU: Intel i7 / AMD Ryzen 7 or better.
RAM: 16 GB or more.
GPU: NVIDIA GTX 1060+ (CUDA 11.0+ compatible).
Disk: SSD with 10 GB+ free space for batch image storage.
Display: 2K or 4K monitor for high-res preview.

—

GPU Acceleration

🧠 Deep Learning & CUDA

While MIRA can run on a CPU, using an NVIDIA GPU significantly improves performance:

CPU-only: ~1-3 frames per second (FPS) for detection.
GPU-accelerated: ~20-30+ frames per second (FPS) for real-time smoothness.

If you have an NVIDIA card, ensure you install the appropriate CUDA Toolkit and cuDNN library to enable hardware acceleration.

—

💡 Connection Tip

Always connect the ArduCam to a USB 3.0 (Blue) port. Using a USB 2.0 port will significantly limit the frame rate and may cause connection stability issues during high-resolution captures.