Karri S P K, Chakraborty Debjani, Chatterjee Jyotirmoy
School of Medical Science and Technology, IIT Kharagpur, Kharagpur, India.
Department of Mathematics, IIT Kharagpur, Kharagpur, India.
Biomed Opt Express. 2017 Jan 4;8(2):579-592. doi: 10.1364/BOE.8.000579. eCollection 2017 Feb 1.
We present an algorithm for identifying retinal pathologies given retinal optical coherence tomography (OCT) images. Our approach fine-tunes a pre-trained convolutional neural network (CNN), GoogLeNet, to improve its prediction capability (compared to random initialization training) and identifies salient responses during prediction to understand learned filter characteristics. We considered a data set containing subjects with diabetic macular edema, or dry age-related macular degeneration, or no pathology. The fine-tuned CNN could effectively identify pathologies in comparison to classical learning. Our algorithm aims to demonstrate that models trained on non-medical images can be fine-tuned for classifying OCT images with limited training data.
我们提出了一种用于在给定视网膜光学相干断层扫描（OCT）图像的情况下识别视网膜病变的算法。我们的方法对预训练的卷积神经网络（CNN）GoogLeNet进行微调，以提高其预测能力（与随机初始化训练相比），并在预测过程中识别显著响应，以了解学习到的滤波器特征。我们考虑了一个包含患有糖尿病性黄斑水肿、干性年龄相关性黄斑变性或无病变的受试者的数据集。与传统学习相比，微调后的CNN能够有效地识别病变。我们的算法旨在证明，在非医学图像上训练的模型可以通过有限的训练数据进行微调，以对OCT图像进行分类。
