Thyroid lesion detection system based on NPU local semantic segmentation

Abstract

The invention discloses a thyroid lesion detection system based on NPU local semantic segmentation. The thyroid lesion detection system comprises a high-resolution thyroid lesion image acquisition module, a neural network semantic segmentation module, an NPU edge device image processing module and a visualization module. The system collects thyroid lesion pictures as a data set; constructing a semantic segmentation network for thyroid lesion identification by using a deep learning framework; marking the original thyroid image to construct a data label, and sending the original image and the label into a neural network for training; converting the trained model, deploying the model into an NPU processor, and writing drive programs of a hardware resource calling module, an image acquisition module, a neural network reasoning module, an image post-processing module and a visual display module of the NPU processor; a camera is used for collecting high-resolution static pictures and storing the pictures locally; and the NPU hardware acceleration module is used for locally segmenting thyroid lesion pictures acquired by the camera through a neural network algorithm and displaying a detection result.

Description

technical field [0001] The invention relates to the field of intelligent hardware and auxiliary medical diagnosis, in particular to a thyroid lesion detection system based on NPU local detection semantic segmentation. Background technique [0002] With the continuous advancement of science and technology, industry, medical care, and life are all developing in the direction of intelligence. The arrival of the era of artificial intelligence has brought great convenience to all aspects of life. Smart medical care is a huge branch of the field of artificial intelligence. . The thyroid is the largest endocrine gland in the human body. It is located in the front of the neck of the human body, below the Adam's apple, and is shaped like a butterfly. Its main function is to synthesize thyroid hormones, regulate the body's metabolism, and play an important role in human growth and metabolism. Its volume is very small, generally only 20-30g, and thyroid nodules appear after abnormal p...

AI Technical Summary

Problems solved by technology

[0003] At present, the diagnosis method of thyroid nodules is to take ultrasound images and diagnose thyroid lesions through the experience of doctors, but this method has many defects, for example, ultrasound images rely on X-ray computed tomography (CT), and the image clarity is not high , the difference between various tissues and organs is not obvious; rely on the accumulation of experience of doctors; shortage of medical resources, long time to see a doctor, etc.

Images