Pulmonary Tuberculosis Edge Diagnosis System Based on MindSpore Framework: Low-cost and High-precision Implementation with Ascend 310 Chip
HaoYu Li
TL;DR
This work tackles the challenge of affordable, accurate pulmonary tuberculosis diagnosis in resource-limited settings by deploying a lightweight deep learning model on edge hardware. It combines MobileNetV3-Large with SE and Swish enhancements, run on Huawei's Ascend 310 chip via the MindSpore framework, to deliver real-time X-ray analysis on an Orange Pi AIPro device. On a test set of 4,148 images, it achieves an accuracy of $99.1\%$ and an AUC of $0.99$, while the total hardware cost remains below $150 and power use stays at about 8 W. This approach enables scalable, low-cost AI-assisted TB screening with potential extensions to multi-task clinical decision support and improved user interfaces for grassroots healthcare deployment.
Abstract
Pulmonary Tuberculosis (PTB) remains a major challenge for global health, especially in areas with poor medical resources, where access to specialized medical knowledge and diagnostic tools is limited. This paper presents an auxiliary diagnosis system for pulmonary tuberculosis based on Huawei MindSpore framework and Ascend310 edge computing chip. Using MobileNetV3 architecture and Softmax cross entropy loss function with momentum optimizer. The system operates with FP16 hybrid accuracy on the Orange pie AIPro (Atlas 200 DK) edge device and performs well. In the test set containing 4148 chest images, the model accuracy reached 99.1\% (AUC = 0.99), and the equipment cost was controlled within \$150, providing affordable AI-assisted diagnosis scheme for primary care.