srch

Loss odyssey in medical image segmentation Med Image Anal. 2021 Jul;71:102035. doi: 10.1016/j.media.2021.102035. Epub 2021 Mar 19. Authors Jun Ma 1 ...

20.03.2021 · New segmentation results (based on nnUNetV2) of different loss functions in "Loss Odyssey in Medical Image Segmentation".

Abstract. Loss functions are one of the crucial ingredients in deep learning-based medical image segmentation methods. Many loss functions have been ...

The loss function is an important component in deep learning-based segmentation methods. Over the past five years, many loss functions have been proposed ...

20.03.2021 · Segmentation results of different loss functions in "Loss Odyssey in Medical Image Segmentation".

01.07.2021 · Dice loss has been widely used in many top solutions of medical image segmentation challenges ( Bernard, Lalande, Zotti, Cervenansky, Yang, Heng, Cetin, Lekadir, Camara, Ballester, et al., 2018, Bakas, Reyes, et Int, Menze ). In addition to cross entropy and Dice loss, more than 10 loss functions have been proposed for segmentation CNNs.

05.05.2021 · New segmentation results (based on nnUNetV2) of different loss functions in "Loss Odyssey in Medical Image Segmentation".

This paper presents a comprehensive review of segmentation loss functions in an organized manner and conducts the first large-scale analysis ...

Loss odyssey in medical image segmentation. Jun Ma, Jianan Chen 0001, Matthew Ng, Rui Huang, Yu Li, Chen Li, Xiaoping Yang, Anne L. Martel. Loss odyssey in ...

Medical Image Analysis. Volume 71, July 2021, 102035. Loss odyssey in medical image segmentation. Author links open overlay panel Jun Ma a Jianan Chen b Matthew Ng b Rui Huang b Yu Li a Chen Li d Xiaoping Yang d Anne L. Martel b c. ... We present the first comprehensive review and comparison of the existing plug-and-play segmentation loss ...

In the training stage of the segmentation network, DSC loss [41] serves as the mainstream loss function that is widely used in medical image segmentation.

Abstract: Loss functions are one of the crucial ingredients in deep learning-based medical image segmentation methods. Many loss functions ...

In particular, IoU loss (also known as soft Jaccard) simply replaces the segmentation in Jaccard index with softmax probabilities, while Lovász loss (Berman et ...

Loss Odyssey in Medical Image Segmentation · Jun Ma · Jianan Chen · Matthew Ng · Rui Huang · Yu Li ...