Information Technology University, Lahore, Pakistan.
Turner Institute for Brain and Mental Health, Monash University, Clayton, Australia.
Med Biol Eng Comput. 2022 Oct;60(10):2797-2811. doi: 10.1007/s11517-022-02633-w. Epub 2022 Jul 20.
In recent years, deep learning (DL) techniques have provided state-of-the-art performance in medical imaging. However, good quality (annotated) medical data is in general hard to find due to the usually high cost of medical images, limited availability of expert annotators (e.g., radiologists), and the amount of time required for annotation. In addition, DL is data-hungry and its training requires extensive computational resources. Furthermore, DL being a black-box method lacks transparency on its inner working and lacks fundamental understanding behind decisions made by the model and consequently, this notion enhances the uncertainty on its predictions. To this end, we address these challenges by proposing a hybrid model, which uses a Bayesian convolutional neural network (BCNN) for uncertainty quantification, and an active learning approach for annotating the unlabeled data. The BCNN is used as a feature descriptor and these features are then used for training a model, in an active learning setting. We evaluate the proposed framework for diabetic retinopathy classification problem and demonstrate state-of-the-art performance in terms of different metrics.
近年来，深度学习（DL）技术在医学成像领域取得了最先进的性能。然而，由于医学图像的成本通常较高、专家注释者（例如放射科医生）的可用性有限以及注释所需的时间，通常很难找到高质量（注释）的医疗数据。此外，DL 需要大量数据，其训练需要大量的计算资源。此外，DL 作为一种黑盒方法，其内部工作缺乏透明度，并且缺乏对模型做出的决策背后的基本理解，因此，这一概念增加了其预测的不确定性。为此，我们通过提出一种混合模型来解决这些挑战，该模型使用贝叶斯卷积神经网络（BCNN）进行不确定性量化，并使用主动学习方法对未标记的数据进行注释。BCNN 用作特征描述符，然后将这些特征用于主动学习环境中的模型训练。我们评估了所提出的用于糖尿病视网膜病变分类问题的框架，并在不同指标方面展示了最先进的性能。
