Lightweight skin disease lesion area segmentation method and system
By designing the LDCM module and TinyDenseUNet model, the efficient segmentation problem of U-Net networks under limited resources is solved, efficient, fast and accurate segmentation of skin disease lesions is achieved, and the learning and representation ability of the model is improved.
CN120431111APending Publication Date: 2025-08-05SUZHOU INST FOR ADVANCED STUDY USTC +1
View PDF 0 Cites 0 Cited by
Patent Information
- Application Number
- CN202510508133.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-22
- Publication Date
- 2025-08-05
Smart Images
Figure CN120431111A_ABST
Abstract
The invention discloses a lightweight skin disease lesion area segmentation method and system. The lightweight skin disease lesion area segmentation method comprises the following steps: acquiring a to-be-segmented dermatoscope image; the dermatoscope image to be segmented is input into a trained image segmentation model TinyDenseUNet, and the trained image segmentation model TinyDenseUNet is used for segmenting the dermatoscope image to be segmented; the TinyDenseUNet model is a U-Net type network and is composed of three parts, namely an encoder, a decoder and a jump connection part; wherein the LDCM module and the down-sampling module form an encoder of the image segmentation model, and the up-sampling module and the LDCM module form a decoder of the image segmentation model; and outputting a final segmentation result. According to the method, efficient, rapid and accurate segmentation can be realized under limited computing resources and data set scale, and the method is suitable for clinical diagnosis and medical research.
Need to check novelty before this filing date? Find Prior Art