2020/03/09 by Olivier Vincent, Vincent, Olivier, Charley Gros +5 · 1 citation
Computer Science · Engineering · Medicine · #AI in cancer detection #Cervical Cancer and HPV Research #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #FOS: Electrical engineering #Image and Video Processing (eess.IV) #Machine Learning (cs.LG) #Medical Imaging and Analysis #Multiple Sclerosis Research Studies #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2003.04377
openalex publication_date 2020/03/09 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Despite recent improvements in medical image segmentation, the ability to\ngeneralize across imaging contrasts remains an open issue. To tackle this\nchallenge, we implement Feature-wise Linear Modulation (FiLM) to leverage\nphysics knowledge within the segmentation model and learn the characteristics\nof each contrast. Interestingly, a well-optimised U-Net reached the same\nperformance as our FiLMed-Unet on a multi-contrast dataset (0.72 of Dice\nscore), which suggests that there is a bottleneck in spinal MS lesion\nsegmentation different from the generalization across varying contrasts. This\nbottleneck likely stems from inter-rater variability, which is estimated at\n0.61 of Dice score in our dataset.\n