Abstract:
Early diagnosis of retinal diseases, the leading cause of permanent blindness, requires segmenting retinal layers in optical coherence tomography (OCT) images. Thus, segmentation is crucial to vision preservation.
Expert ophthalmologists must manually segment retinal layers due to a lack of practical automated methods. This paper presents a feature-learning regression network-based OCT image segmentation method without human bias.
The deep neural network regression predicts the retinal boundary pixel from an image segment’s intensity, gradient, and adaptive normalized intensity score (ANIS). The computational complexity analysis shows that reformulating segmentation as a regression problem eliminates the need for a huge dataset and reduces complexity.
ANIS helps the method work accurately and quickly on OCT images with intensity variances, low-contrast regions, speckle noise, and blood vessels. In 114 images, identifying eight boundaries took 10.596 s per image, and training each boundary line took 30 s.
The accuracy Dice similarity coefficient was 0.966. The average absolute pixel distance of manual and automatic segmentation using the proposed scheme was 0.612.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
