top of page

Edge AI Platform for Dermatology Diagnostic Device 

Client Background

Our client, a prominent leader in the medical technology sector, aimed to revolutionize skin cancer detection with a handheld diagnostic device. Focused on innovation, they sought a solution capable of high-resolution skin imaging and Edge AI-driven diagnostics. Regami collaborated to address this challenge, ultimately delivering a tailored embedded system that significantly advanced dermatological diagnostics.


Challenge

Our client, a leader in the field of medical technology, embarked on a mission to develop a handheld device for skin cancer detection. The challenge was clear - they sought a solution that could capture high-resolution, close-up images of the skin and apply Edge AI-driven diagnostics to identify potential skin cancer. The client required a vision-integrated board assembly housing an AI-capable processor running Linux to accommodate model training with the vast dataset of skin images collected.

MicrosoftTeams-image (2).png

Reach out to us!

Let’s bring your ideas to life

Solution

Regami collaborated closely with the client to craft a tailored solution. We delivered a comprehensive board-level embedded system that met their precise requirements:


  • 4K Color Imager with NIR Sensitivity: We integrated a high-resolution 4K color imager with Near-Infrared (NIR) sensitivity to capture high-quality images in various lighting conditions, ensuring the device's reliability.


  • AI-Enabled Processor: The system featured a Cortex A73 processor equipped with a built-in AI coprocessor that supported TensorFlow models. This allowed for efficient real-time analysis of captured skin images for potential cancer detection.


  • Optimized Memory and Storage: Regami designed the system with the necessary memory and storage to handle the vast dataset required for model training and image storage, ensuring seamless performance.


  • High-Performance Optics: We selected the optimal optics for the imager, guaranteeing superior image quality for diagnostic purposes.


  • Yocto Linux Integration: Our team implemented Yocto Linux, enabling the client's software team to seamlessly integrate their custom software and train AI models.

factory-worker-operating-drilling-cnc-ma