Related Experiment Video
Updated: Aug 27, 2025
05:56
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
Published on: April 14, 2023
2.6K
An Improved Image Classification Method for Cervical Precancerous Lesions Based on ShuffleNet.
Shan Fang1, Jiahui Yang1, Minghui Wang1
1College of Quality and Technical Supervision, Hebei University, Baoding 071002, China.
Computational Intelligence and Neuroscience
|September 23, 2022
Summary
This study introduces an improved ShuffleNet deep learning model for classifying cervical precancerous lesions from colposcopy images. The AI model achieves high accuracy and efficiency, aiding clinical screening.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oncology
Background:
- Deep learning aids clinical screening for cervical precancerous lesions.
- Existing algorithms struggle with simultaneous high accuracy and speed.
- Accurate classification is crucial for timely intervention.
Purpose of the Study:
- To develop a high-accuracy, fast cervical precancerous lesion classification method.
- To enhance ShuffleNet performance using channel attention.
- To address limitations in existing deep learning models for cervical lesion detection.
Main Methods:
- A ShuffleNet-based convolutional neural network (CNN) with channel attention was developed.
- Colposcopy images were augmented to address data scarcity and imbalance.
- The dataset included five categories: normal, cervical cancer, LSIL (CIN1), HSIL (CIN2/CIN3), and cervical neoplasm.
Main Results:
- The proposed CNN model achieved test accuracies of 81.23% and 81.38%.
- The classifier obtained an Area Under the Curve (AUC) score of 0.99.
- The AI network demonstrated strong performance in classification accuracy and model size.
Conclusions:
- The AI-powered colposcopy image classification network shows high clinical applicability.
- The ShuffleNet-based model offers a promising solution for efficient and accurate cervical precancerous lesion detection.
- The study highlights the potential of deep learning in improving cervical cancer screening.
