Performance Analysis of Various Deep Learning Denoise Techniques on Retinal OCT Images

Optical Coherence Tomography (OCT) is increasingly utilized to detect eye problems using retinal structure data. Many retinal problems are diagnosed and monitored by OCT. Acquisition speckle noise weakens OCT images. OCT images are typically distorted by speckle noise, making diagnosis and interpretation challenging. Speckle noise sharpens layer and blood vessel borders in OCT images. OCT image creation is like ultrasound imaging, except for the imaging mode. OCT images are taken using light beams, unlike ultrasound. Age affects many things. These are varied. Age-related Macular Degeneration (AMD) attacks the retina and macula, which are important to vision, in those over 50. Retina's Pigment Epithelium (RPE) anomalies and yellow drusen deposits suggest retinal degeneration. Speckle noise may impair the diagnostic accuracy of high-resolution optical coherence tomography (OCT). Preventing irreversible visual loss requires early AMD detection. All performance metrics show that the Residual Image Denoising Network (RIDN) is the best OCT data denoising approach.