Deep Learning Methods for Retinal Blood Vessel Image Extraction.

      Аннотация: The changes in the retinal blood vessel features are used to diagnose multiple diseases and detect them at an early stage. In order to detect these changes a very crucial step is involved which is called as vessel segmentation process. Different diseases affect retinal vessels like neovascularization in diabetic retinopathy, heart stroke, and cardiovascular diseases. Retina image analysis plays a vital role in the medical diagnosis of multiple diseases. Diabetic retinopathy is considered one of the dreadful diseases that damage the retina and leads to vision loss. So segmentation of these blood vessels guides the ophthalmologists to assess the disease intensity accurately. It was observed that retina blood vessel structures are treated as an indispensable element for ophthalmologists in assessing cardiovascular diseases hence there is a great need to segment these vessel structures accurately which can be accomplished only through computer-aided diagnostic algorithms. In this book, an attempt is made to fulfill the need for accurate, robust, and fast automated vessel segmentation methods by investigating different works of image segmentation algorithms and techniques.

Ключевые слова: Deep Learning, Diabetic retinopathy, retinal blood vessel images, Convolutional neural networks, Biomedical image processing, Image segmentation, deep belief network, Monarch Butterfly Optimization, Distance-based Monarch Butterfly Optimization, Contrast Limited Adaptive Histogram Equalization, Adaptive Weighted Contrast Adjustment