| Skin cancer and various skin-related skin diseases are seriously threatening human health.In most cases,when skin cancer patients are diagnosed,they are already in the advanced stage of skin cancer.At this time,the survival rate of patients is very low because skin cancer cells are extremely susceptible to lesions and easily metastasize.Therefore,the timely diagnosis of skin surface lesions becomes crucial.In the medical field,there are more and more cases where artificial intelligence is combined with clinical medicine for condition diagnosis,so it is important to analyze and study the medical images of skin cancer collected by dermatoscopy through computer.The main research contents of this paper are.（1）Through reading and analyzing a lot of related literature,we found that the skin cancer medical image segmentation based on traditional machine learning algorithm has problems such as too tedious pre-processing work and poor segmentation effect,which cannot effectively assist doctors in diagnosis,and determined the research direction of this paper as skin cancer medical image segmentation algorithm based on deep learning.（2）We designed and built a deep learning-based skin cancer medical image segmentation framework,which includes skin cancer medical image acquisition module,image training module and image detection module.In the image training module,three deep learning image segmentation networks,U-Net,Seg Net and DeepLab V3+,are designed firstly,and the segmentation effects of the three networks in skin cancer medical image segmentation are compared through experiments.（3）Through the above comparative analysis,the improved deep learning segmentation algorithm based on DeepLab V3+ network is proposed for the problems reflected by the three networks in the experiments.The DeepLab V3+ algorithm is improved by adding Transformer module and other methods,and finally the experimental results of the improved Deeo Lab V3+algorithm are compared and analyzed with DeepLab V3+.The experiments prove that the improved DeepLab V3+ segmentation algorithm is better in boundary detail segmentation of medical images of skin cancer compared with DeepLab V3+ segmentation algorithm,meanwhile,the m IOU of the improved DeepLab V3+ segmentation algorithm is improved by1% compared with DeepLab V3+ segmentation algorithm. |
