偏狂症
发表于 2025-3-23 11:43:41
On the Pitfalls of Entropy-Based Uncertainty for Multi-class Semi-supervised Segmentationed solution on a challenging multi-class segmentation dataset and in two well-known uncertainty-based segmentation methods. The reported results demonstrate that by simply replacing the mechanism used to compute the uncertainty, our proposed solution brings consistent improvements.
Conjuction
发表于 2025-3-23 15:37:01
What Do Untargeted Adversarial Examples Reveal in Medical Image Segmentation?sults for uncertain region findings on medical image datasets while only requiring one extra inference from a pre-trained model and short iteration of attack. We expect our novel findings can provide insights for future medical image segmentation problems where detection of subtle variations (e.g.,
Functional
发表于 2025-3-23 21:51:22
Improved Post-hoc Probability Calibration for Out-of-Domain MRI Segmentationodel is tested on cardiac MRI segmentation datasets that contain unseen imaging artifacts and images from an unseen imaging protocol. We demonstrate reduced calibration errors compared with the state-of-the-art calibration algorithm.
官僚统治
发表于 2025-3-24 01:30:13
http://reply.papertrans.cn/95/9412/941132/941132_14.png
Trigger-Point
发表于 2025-3-24 03:51:27
http://reply.papertrans.cn/95/9412/941132/941132_15.png
AGONY
发表于 2025-3-24 06:49:34
Joint Paraspinal Muscle Segmentation and Inter-rater Labeling Variability Prediction with Multi-taskdicting inter-rater labeling variability visualized using a variance map of three raters’ annotations. Our technique is validated on MRIs of paraspinal muscles at four different disc levels from 118 LBP patients. Benefiting from the transformer mechanism and convolution neural networks, our algorith
cochlea
发表于 2025-3-24 11:31:59
Information Gain Sampling for Active Learning in Medical Image Classificationes including the diversity based CoreSet and uncertainty based maximum entropy sampling. Specifically, AEIG achieves . of overall performance with only 19% of the training data, while other active learning approaches require around 25%. We show that, by careful design choices, our model can be integ
interpose
发表于 2025-3-24 15:31:14
http://reply.papertrans.cn/95/9412/941132/941132_18.png
彻底明白
发表于 2025-3-24 21:16:57
http://reply.papertrans.cn/95/9412/941132/941132_19.png
Simulate
发表于 2025-3-25 00:00:40
http://reply.papertrans.cn/95/9412/941132/941132_20.png