E-POSTER DETAIL

Title
“The Doctor as Developer” Building a Fundus Charting App with Gemini Pro
Authors
Manirattan Singh Brar, Anand Singh Brar
Presenting
Manirattan Singh Brar
PURPOSE:
To evaluate the efficacy of a 'no-code' workflow using Gemini Pro to develop a browser-based fundus charting application.
METHODS:
Gemini Pro acted as the primary coding engine. Natural language prompts were used to generate HTML5 and JavaScript code for an interactive HTML Canvas interface. The workflow utilized iterative prompting to debug and refine features (e.g., freehand drawing, standard retinal color codes) without manual programming. Outcome measures were development duration and System Usability Scale (SUS) to assess user interface satisfaction.
RESULTS:
A fully functional, deployable web application was completed in 12 hours. The application successfully supported multi-layer drawing for retinal pathology. Pilot testing yielded a good SUS score of 74/100.
CONCLUSIONS:
Generative AI successfully automates the creation of specialized ophthalmic software. This 'no-code' paradigm significantly reduces technical and financial barriers, allowing clinicians to build bespoke digital tools.