基于多金字塔、局部二值模式（LBP）和方向梯度直方图（HOG）描述符的频域光学相干断层扫描（SD-OCT）图像分类：在糖尿病性黄斑水肿（DME）检测中的应用

Classification of SD-OCT volumes with multi pyramids, LBP and HOG descriptors: application to DME detections.

作者信息

Alsaih Khaled, Lemaitre Guillaume, Vall Join Massich, Rastgoo Mojdeh, Sidibe Desire, Wong Tien Y, Lamoureux Ecosse, Milea Dan, Cheung Carol Y, Meriaudeau Fabrice

出版信息

Annu Int Conf IEEE Eng Med Biol Soc. 2016 Aug;2016:1344-1347. doi: 10.1109/EMBC.2016.7590956.

DOI:10.1109/EMBC.2016.7590956

PMID:28268574

Abstract

This paper deals with the automated detection of Diabetic Macular Edema (DME) on Optical Coherence Tomography (OCT) volumes. Our method considers a generic classification pipeline with preprocessing for noise removal and flattening of each B-Scan. Features such as Histogram of Oriented Gradients (HOG) and Local Binary Patterns (LBP) are extracted and combined to create a set of different feature vectors which are fed to a linear-Support Vector Machines (SVM) Classifier. Experimental results show a promising sensitivity/specificity of 0.75/0.87 on a challenging dataset.

摘要

本文探讨了在光学相干断层扫描（OCT）容积上自动检测糖尿病性黄斑水肿（DME）的方法。我们的方法采用了一个通用的分类流程，包括对每个B扫描进行噪声去除和平坦化的预处理。提取诸如定向梯度直方图（HOG）和局部二值模式（LBP）等特征并将其组合，以创建一组不同的特征向量，这些特征向量被输入到线性支持向量机（SVM）分类器中。实验结果表明，在一个具有挑战性的数据集上，灵敏度/特异性达到了0.75/0.87，前景可观。