Skin Lesion Area Segmentation Using Attention Squeeze U-Net for Embedded Devices.

"the web page https://paperswithcode com/sota/lesion-segmentation-on-isic-2018 contains a list of papers about lesion segmentation on isic 2018 with the source code available.; the complete source code for all the four networks written using tensorflow 2 and python 3 is publicly available at: https://github com/apennisi/att_squeeze_unet table 1 shows the segmentation results obtained on the isic 2017 test set by using the dice similarity coefficient and the threshold jaccard index as quality metrics."

"Declarations Ethics ApprovalThis research study was conducted retrospectively from data obtained for clinical purposes. We used only data coming from publicly available datasets. In particular, we used the ISIC 2017 and 2018 Challenge datasets and the PH2 dataset. Competing InterestsThe authors declare no competing interests. Competing Interests The authors declare no competing interests."