Retinoscan
Development of a portable fundus imaging device with integrated machine learning for automated Diabetic Retinopathy detection
Diabetes mellitus and diabetic retinopathy (DR) present significant global public health challenges. Currently, 10.5% of the world's adult population has diabetes, as reported by the International Diabetes Federation. DR is a leading cause of vision impairment and blindness, affecting 30% to 40% of diabetic individuals. A meta-analysis estimates that 103 million people currently suffer from DR, a figure projected to rise to 161 million by 2045. Early detection is crucial, as timely interventions can prevent up to 95% of vision loss cases. Unfortunately, access to affordable healthcare and specialized eye care services is limited in many low- and middle-income countries, where trained ophthalmologists and advanced imaging tools are often concentrated in urban areas.
My project addresses disparities in diabetic retinopathy (DR) screening by developing a portable fundus imaging device that integrates machine learning for automated DR detection. This innovative device captures high-quality retinal images and utilizes advanced deep learning algorithms to accurately classify the severity of DR. By training on a comprehensive dataset of annotated images, the project ensures high diagnostic accuracy and reliability. The goal is to enhance accessibility and efficiency in DR screening, particularly in underserved areas, thereby facilitating timely diagnosis.
The innovation in this research lies in combining portable retinal imaging technology with state-of-the-art deep learning algorithms for DR detection with a portable device designed for use in diverse environments, from clinics to remote areas
-
Paper presentation at ACMSEGA 2024, an International Conference on Advances inï‚· Communication, Medical Electronics and Smart Grid Automation
-
Among the youngest participants to have a research paper selected for publication ina Scopus-indexed journal.
Study of Biomarker-based approaches for early detection of ALzheimer'susing Machine Learning
This research compares biomarker-based approaches for early Alzheimer’s detection
using machine learning.
Project on Pneumonia Detection using Chest X-Rays
Designed a medical imaging project for pneumonia detection. The machine learning algorithm
could analyze chest X-ray images and accurately classify patients as pneumonic or healthy.
Leveraged image preprocessing and convolutional neural networks (CNNs) to enhance
diagnostic precision.
AquaNova
Developed at ChangeMakers Summer Camp 2025 at Indian Institute of Technology(IIT),Delhi
Delhi thrives on Yamuna,yet it is drowning in pollution endangering the city's health,culture and ecology.
High influx of untreated municipal, industrial,and agricultural wastewater particularly near Najafgarh,Barapullah,Shahdara
Prototype has a two-stage system: Prefiltration for TSS removal, followed by Chlorella. Vulgaris photobioreactor for phosphate and nitrate assimilation under photoautotrophic conditions.
.jpeg)
.jpeg)
.jpeg)
EcoCell
As the Founder of the Science and Sustainability Club at my school and as a quarterfinalist at the
Technovation Girls Challenge 2023, I pioneered the development of an innovative app, Ecocell, designed
to encourage the recycling of lithium-ion batteries, particularly among owners of electric vehicles (EVs).
Recognizing the growing need for responsible waste management in the era of electric vehicles, I
partnered with Attero, a company dedicated to building a circular economy by recovering critical
minerals like lithium, cobalt, and nickel. With their support, our initiative successfully recycled 2 tons of
lithium-ion batteries, demonstrating the impact a simple yet effective idea could have on reducing waste and conserving valuable resources. This initiative not only provided a practical solution to a pressing environmental issue but also helped establish the foundation for future projects aimed at promoting sustainability and responsible resource use.