What is it about?
Lesion detection technology based on deep learning requires huge amounts of high-quality medical image data, but the data from social IoT has the problems of uneven quality and lack of lesion labeling. This paper proposes a core data extraction method for multi-class lesion detection, to effectively reduce the labeling burden.
Featured Image
Photo by Accuray on Unsplash
Why is it important?
Compared with serial and parallel extraction of multiple single-class core data, our method greatly improves the extraction speed and reduces the resource waste of multi-model training. And our experimental results show that our method can effectively extract core data of multiple lesions from low-quality medical images, and realize higher accuracy and better interpretability.
Perspectives
We hope that by cooperating with hospitals, we can expand the data set from real cases, improve models by applying them to various diseases, and provide effective solutions to assist doctors in diagnosis.
Zhu Feihong
Central South University
Read the Original
This page is a summary of: LesionTalk: Core Data Extraction and Multi-class Lesion Detection in IoT-based Intelligent Healthcare, ACM Transactions on Sensor Networks, March 2023, ACM (Association for Computing Machinery),
DOI: 10.1145/3526194.
You can read the full text:
Contributors
The following have contributed to this page
