Growing up, Tisha Sehrawat lived near lakes that were never supposed to be touched by fire. The job of these water bodies around Bangalore, constructed centuries ago, was to keep the monsoon overflow at bay while simultaneously cool the city. . Fast forward to today, several lakes heat up and burn, suffocated by irresponsible waste dumping that has been going on for decades It was on the banks of one of these lakes that Sehrawat, now a 12th grader at 10X International School, asked herself a question that would make her think for years to come: where does all that plastic actually go?
With the passing of time, she learned that it never really goes away. It breaks down into microplastics, into tiny fragments that slip into rivers, soil and eventually the food chain, increasing in a concentric manner until they reach people. Recent studies have detected microplastics in human blood, lung tissue and even the placenta, and researchers are yet to understand what that means for long-term health.This uncharted gap was the real problem for Sehrawat. The benchmark detection tools FTIR and Raman spectroscopy, cost tens of thousands of rupees and require a trained spectroscopist, making testing unaffordable for most clinics, schools and households.
“I kept coming back to the same question: why does something that might already be inside our bodies stay invisible to the people trying to protect us?” Sehrawat said. “I didn’t want to just study pollution. I wanted to build something that put the power to check for it back in people’s hands, whether that’s a clinic, a school or a home.”
To answer that question, a Raspberry Pi, a microscope camera module and a sealed, controlled illumination imaging chamber were used to create PlastiXcan, a portable micro-plastic sensor. It costs less than Rs. 4000 to construct and use on-device computer vision and machine learning to categorise fibers, fragments and beads by shape and texture. It also produces particle counts and size estimations without needing a reagent or chemical digestion phase. Sehrawat has tested it on more than 500 household water and milk samples, and has found 98 percent detection accuracy in water and 94 percent in milk, as well as a 95 percent reduction in the cost of testing and an 87 percent gain in testing speed compared with lab-based spectroscopy. Sehrawat is currently discussing large-scale contamination monitoring with the Food Safety and Standards Authority of India.
The work has now successfully passed a bureaucratic hurdle. The Indian Patent Office published Sehrawat’s patent application for a “system and method for non-destructive microplastic identification in food and water using controlled optical imaging and edge-based artificial intelligence” on August 14, 2026. The filing describes an embedded pipeline that isolates particle images, rejects artifacts and categorizes plastic versus non-plastic material on-device, without requiring any external computing or lab-grade spectroscopy .
In Sehrawat’s work, PlastiXcan fits into a larger theme that looks at an issue from different perspectives. She has noticed a trend that the easiest issues to resolve before they get to a public hospital are those related to public health. She has studied prion disease, shadowed medics at Fortis and Narayana Health, and worked with patient intake at a community health clinic. To ease patient/caregiver anxiety, she has also created a mobile app. . The subject of her TEDx talk “Fear Is the First Symptom.” “I’m interested in the moment right before someone becomes a patient,” she said. “With PlastiXcan, that means catching a disease before it reaches someone’s plate instead of treating the illness later on. The most useful place for me to be working at, therefore, is if I can help shift the intervention upstream of the hospital, and prevent the outbreak of any disease altogether.”