IEEE

A Novel Transformer-CNN Hybrid Deep Learning Architecture for Robust Broad-Coverage Diagnosis of Eye Diseases on Color Fundus Images

Abstract: We present a novel and robust deep-learning architecture that takes into account the pathological characteristics of eye diseases on color fundus images. The proposed hybrid architecture is ...
IEEE

MSSEAG-UNet: A Novel Deep Learning Architecture for Cloud Segmentation in Fisheye Sky Images and Solar Energy Forecast

Abstract: This study analyzes sky images captured using a ground-based fisheye camera, aiming to address the challenge of accurately segmenting clouds, which is difficult due to their fuzzy and ...
Nature

Machine-learning model generates images using light

Machine-learning models identify relationships in a data set (called the training data set) and use this training to perform operations on data that the model has not encountered before. This could ...