Wellcome Centre for Integrative Neuroimaging, FMRIB, Nuffield Department of Clinical Neurosciences, University of Oxford, UK; Oxford Machine Learning in NeuroImaging Lab (OMNI), Department of Computer Science, University of Oxford, UK.
Wellcome Centre for Integrative Neuroimaging, FMRIB, Nuffield Department of Clinical Neurosciences, University of Oxford, UK; Australian Institute for Machine Learning (AIML), School of Computer Science, University of Adelaide, Adelaide, Australia; South Australian Health and Medical Research Institute (SAHMRI), North Terrace, Adelaide, Australia.
Med Image Anal. 2022 Oct;81:102583. doi: 10.1016/j.media.2022.102583. Epub 2022 Aug 17.
Acquisition of high quality manual annotations is vital for the development of segmentation algorithms. However, to create them we require a substantial amount of expert time and knowledge. Large numbers of labels are required to train convolutional neural networks due to the vast number of parameters that must be learned in the optimisation process. Here, we develop the STAMP algorithm to allow the simultaneous training and pruning of a UNet architecture for medical image segmentation with targeted channelwise dropout to make the network robust to the pruning. We demonstrate the technique across segmentation tasks and imaging modalities. It is then shown that, through online pruning, we are able to train networks to have much higher performance than the equivalent standard UNet models while reducing their size by more than 85% in terms of parameters. This has the potential to allow networks to be directly trained on datasets where very low numbers of labels are available.
获取高质量的人工标注对于分割算法的发展至关重要。但是，为了创建它们，我们需要大量的专家时间和知识。由于在优化过程中必须学习大量参数，因此需要大量标签来训练卷积神经网络。在这里，我们开发了 STAMP 算法，允许同时训练和修剪 UNet 架构，用于医学图像分割，并针对通道进行有针对性的丢弃，以使网络对修剪具有鲁棒性。我们在分割任务和成像模式中展示了该技术。然后表明，通过在线修剪，我们能够训练网络，使其性能远高于等效的标准 UNet 模型，同时将其参数减少 85%以上。这有可能允许网络直接在标签数量非常少的数据集上进行训练。
