Proceedings of the 2022 3rd International Conference on Artificial Intelligence and Education (IC-ICAIE 2022)

<Previous Article In Volume
Next Article In Volume>

Study on Improving Model Accuracy by Using Non-medical Image Pretraining Set based on ImageNet Weight

Authors
Xiangwu Guo1, 1, *
1Software Engineering, South China University of Technology, Panyu District, Guangzhou, China
*Corresponding author. Email: sexiangwu_guo@mail.scut.edu.cn
Corresponding Author
Xiangwu Guo
Available Online 27 December 2022.
DOI
10.2991/978-94-6463-040-4_122How to use a DOI?
Keywords
MobileNetV2; Transfer Learning; Skin cancer; ImageNet
Abstract

Although modern medical systems have relatively complete skin cancer detection methods, it is still very difficult to detect early skin cancer themselves. Therefore, this paper intends to propose a lightweight model network for patient self-detection of skin cancer. In order to implement a lightweight model, this article chose MoblieNetV2 as part of the model construction. Faced with the inherent difficulty of insufficient in-depth training of lightweight models, this paper tries to use transfer learning to improve model accuracy. Since there is no pre-training set model for skin cancer detection, this paper boldly used ImageNet natural image and training set model to achieve the purpose of improving model accuracy and reducing loss. Through the control experiment set under the same conditions, it is found that the accuracy of the test set using Imagenet can reach 88%, which is very improved compared with the performance without Imagenet. The experimental results clearly show that it is feasible to use natural images and medical image detection. Therefore, this paper can try to apply more natural images to medical image detection to make up for the lack of medical image pre-training set.

Copyright
© 2023 The Author(s)
Open Access
Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

<Previous Article In Volume
Next Article In Volume>
Volume Title
Proceedings of the 2022 3rd International Conference on Artificial Intelligence and Education (IC-ICAIE 2022)
Series
Atlantis Highlights in Computer Sciences
Publication Date
27 December 2022
ISBN
10.2991/978-94-6463-040-4_122
ISSN
2589-4900
DOI
10.2991/978-94-6463-040-4_122How to use a DOI?
Copyright
© 2023 The Author(s)
Open Access
Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Cite this article

risenwbib
TY  - CONF
AU  - Xiangwu Guo
PY  - 2022
DA  - 2022/12/27
TI  - Study on Improving Model Accuracy by Using Non-medical Image Pretraining Set based on ImageNet Weight
BT  - Proceedings of the 2022 3rd International Conference on Artificial Intelligence and Education (IC-ICAIE 2022)
PB  - Atlantis Press
SP  - 804
EP  - 809
SN  - 2589-4900
UR  - https://doi.org/10.2991/978-94-6463-040-4_122
DO  - 10.2991/978-94-6463-040-4_122
ID  - Guo2022
ER  -