Computer Vision for Real-Time Monkeypox Diagnosis on Embedded Systems

TL;DR

This work addresses the need for rapid, low-resource mpox diagnosis by deploying an edge AI tool on the Jetson Orin Nano using MobileNetV2 with transfer learning on Monkeypox skin lesion datasets. It leverages NVIDIA TensorRT mixed-precision quantization (FP32/FP16/INT8) to compress the model while preserving accuracy, achieving an F1-Score of $93.07\%$ and notable reductions in size, latency, and power consumption. The system is accessible via a Wi-Fi AP hotspot with a web-based interface for uploading images and viewing predictions, enabling practical use in underserved areas. Overall, the study demonstrates a scalable, energy-efficient edge-diagnostic solution that can inform broader deployment of edge-AI medical tools in low-resource settings.

Abstract

The rapid diagnosis of infectious diseases, such as monkeypox, is crucial for effective containment and treatment, particularly in resource-constrained environments. This study presents an AI-driven diagnostic tool developed for deployment on the NVIDIA Jetson Orin Nano, leveraging the pre-trained MobileNetV2 architecture for binary classification. The model was trained on the open-source Monkeypox Skin Lesion Dataset, achieving a 93.07% F1-Score, which reflects a well-balanced performance in precision and recall. To optimize the model, the TensorRT framework was used to accelerate inference for FP32 and to perform post-training quantization for FP16 and INT8 formats. TensorRT's mixed-precision capabilities enabled these optimizations, which reduced the model size, increased inference speed, and lowered power consumption by approximately a factor of two, all while maintaining the original accuracy. Power consumption analysis confirmed that the optimized models used significantly less energy during inference, reinforcing their suitability for deployment in resource-constrained environments. The system was deployed with a Wi-Fi Access Point (AP) hotspot and a web-based interface, enabling users to upload and analyze images directly through connected devices such as mobile phones. This setup ensures simple access and seamless connectivity, making the tool practical for real-world applications. These advancements position the diagnostic tool as an efficient, scalable, and energy-conscious solution to address diagnosis challenges in underserved regions, paving the way for broader adoption in low-resource healthcare settings.

Figures (6)

• Figure 1: Web Interface
• Figure 2: Confusion Matrix for Held-Out Dataset
• Figure 3: Post-Compression Model Size
• Figure 4: Post-Compression Average Inference Time
• Figure 5: Post-Compression Throughput