One of the most remarkable domains of generative AI technologies such as ChatGPT currently revolves around improved computer vision. AI is capable of comprehending and interpreting information from images. This is the reason we now have highly sophisticated image and video creation models in ChatGPT, Gemini, Firefly, and other AI applications.
Models such as ChatGPT o3 can precisely determine the location of an image by examining its features. Google provides advanced photo editing capabilities within its Photos application, as well as directly in Gemini. These features enable you to modify real photographs in ways that were previously unachievable.
These AI abilities related to images aren’t solely employed to create memes or overload OpenAI’s servers. Scientists are crafting AI models that can understand images for various applications, including healthcare.
The most recent research highlighting such progress originates from China.
Scientists from various universities have successfully identified a person’s age with remarkable accuracy by having AI analyze an image of their retina. The findings also indicated disparities between the individual’s age and the age of the eye. The researchers discovered that the age discrepancy provided by the AI can be particularly advantageous for women. A straightforward retinal scan could assist healthcare providers in offering improved support to couples seeking to conceive and to women at risk for early menopause.
Retinal fundus imaging, or an image of the back of the eye, allows doctors to observe microvascular characteristics that reflect systemic aging. An AI developed with thousands of images can then forecast the age of the eye and compare it to the individual’s actual age to “forecast retinal age from fundus images with high accuracy.”
The researchers employed an AI known as Frozen and Learning Ensemble Crossover (FLEX) to estimate retinal age from fundus images. They trained FLEX using over 20,000 eye photographs from more than 10,000 adults across various age groups to teach it how the back of the eye appears as people age. FLEX also examined more than 2,500 images from nearly 1,300 pre-menopausal women.
