Catscan
HarmoNeo, conceptualized by Subhro Dutta, is a modern electronic harmonium blending traditional acoustics with digital enhancements. It features a microcontroller-driven system that mimics bellow movements using sensors, offering adjustable sound dynamics.
With a cost-effective design, it aims to provide affordability, portability, and enhanced musical flexibility to harmonium enthusiasts.
With a cost-effective design, it aims to provide affordability, portability, and enhanced musical flexibility to harmonium enthusiasts.
Catscan Work Journal
Cat-Scan couples smartphone ocular imaging with a pupil-first AI engine for rapid, local cataract pre-screening.
Designed for low-resource settings, it streamlines triage and referrals—cutting travel and accelerating care.
Designed for low-resource settings, it streamlines triage and referrals—cutting travel and accelerating care.
MAY DIARY
Identifies the rural cataract screening gap and defines the goal: a low-cost AI app for village use. Assigns roles, confirms clinical partner, and sets success metrics.
LEARN MOREJUNE DIARY
Curates and cleans eye images. Adopts a pupil-first pipeline (upload → adjust pupil → extract → predict) to focus the model on the right signal.
LEARN MOREJULY DIARY
Builds a lightweight mobile prototype for low-connectivity settings. Tests simple screens and visit logging with feedback from ophthalmologists.
LEARN MOREAUGUST DIARY
Connects image processing to the model interface and finalizes basic database flows. Drafts a small field-trial plan and readiness checklist for on-device predictions.
LEARN MORESeptember Diary
In September, the Cat-Scan team focused on training and testing a Convolutional Neural Network (CNN) model for cataract detection using categorized pupil images. Data collection, AI modeling, and doctor validation sessions marked the transition from app development to real-world machine learning implementation
LEARN MOREOctober Diary
In October, the Cat-Scan team finalized their dissertation, adding visuals, screenshots, and clear explanations of the AI model and business problem. They also prepared for the final presentation, integrating feedback and aligning the report with real-world outcomes.
LEARN MORE