Created on 11th February 2024
•
Have you ever stopped to wonder what's really in your food? Picture this: you grab a protein bar or snack from your morning swag bag, but do you truly know what's fueling your body? The truth is, that most of us blindly trust food labels without realizing the hidden ingredients and marketing gimmicks lurking within. We've all heard the age-old adage, 'biscuits are better than chips or chocolates,' but did you know that biscuits can pack more sugar than your favorite chocolate bar?
Enter SwasthkAI – your ultimate solution to decode food labels and make informed choices. With SwasthkAI, you'll never fall prey to misleading marketing tactics again. Our app empowers you to scan food packaging, uncover detailed ingredient data, and assess the health impacts based on your individual preferences – whether you're managing diabetes, allergies, or other health conditions.
Gone are the days of blindly trusting brands. With SwasthkAI, you're in control of your health journey. So, next time you reach for a snack, make it an informed choice with SwasthkAI by your side."
Problem it solves: SwasthkAI revolutionizes the way we approach food consumption by empowering users to scan food packaging, extract detailed ingredient data, and assess health impacts based on individual health preferences. Say goodbye to marketing gimmicks and hello to informed choices tailored for diabetics, patients, and anyone seeking healthier eating habits. Whether you're deciphering labels or avoiding allergens, SwasthkAI is your trusted companion for a healthier lifestyle.
During the project development, we encountered several hurdles. Firstly, locating a suitable dataset containing Indian food ingredients proved to be challenging. Despite extensive efforts, we couldn't find a dataset with the required information. Consequently, we decided to collect the data beforehand and annotated it using Roboflow.
Secondly, determining the precise scope of our project was another challenge. We had to carefully consider which specific ICP (Ideal Customer Profile) we wanted to target.
Lastly, integrating our machine learning models, preprocessing techniques, enhancements, and APIs into a cohesive frontend posed a significant challenge. With the guidance of mentors, we opted to develop our frontend using Streamlit, which helped streamline the process.
Tracks Applied (2)
Major League Hacking
Technologies used