Geetha Manjunath says her experience in working with two American technology companies, based out of Bengaluru, is helping her in the journey as an entrepreneur. She is a hard core technology person and working in the India labs of the two US companies helped her further hone her technical skills. She did not have to worry about the commercial aspect of any project that the labs took up, though there was always the need to prove the commercial use and viability of any project. Now, as an entrepreneur and founder of Niramai Health Analytix, a deep-tech start-up that seeks to find automated solutions for critical health problems, she has to keep in mind the commercial issue too.

During her corporate career, she worked on a number of India-centric projects, especially those with a larger social impact, including in healthcare and transportation. One such was in using thermal imaging, Machine Learning and Artificial Intelligence to screen for breast cancer. When the US company she was working with decided to close down all exploratory activities in India, Geetha says she and a few others decided to quit the company and buy the patent from the US company and start their own venture.

Accurate diagnosis

According to her, they had done field trials using high-resolution thermal cameras and found they were accurate in their diagnosis, as high as 90-95 per cent. Geetha and her team, which includes co-founder Nidhi Mathur, incorporated Niramai in July 2016 and all of them quit their job with the American company in January 2017. They needed to raise funds to pay for the patents. That is when they came across Manish Singhal of pi Ventures, who was ready to sign a term sheet in just about a month of getting Niramai’s proposal. There were other investors too and Niramai was off the ground.


emerging chart


“Breast cancer,” says Geetha, “is the largest killer in women. Nearly 75,000 women die every year in India of breast cancer. The unfortunate part is, 90-95 per cent of the deaths can be prevented through screening.”

The thermal cameras are imported from Sweden, off-the-shelf, while Niramai has developed the software and Machine Learning and Artificial Intelligence capabilities that help in the detection. The procedure is totally non-invasive, says Geetha, adding that a two-minute thermal image of a woman’s breast is taken and the temperature distribution on the chest measured. A few personal details such as family history are taken down. The image has different colours representing different temperatures. That is analysed using image processing and Artificial Intelligence.

“We extract 117 thermal parameters. We pass it through a probability model to say whether there is a chance of the woman getting cancer or not. That we call thermalytix – thermal analytics,” says Geetha. Niramai gives a health score for all the women who come to it for screening.

Niramai sells the hardware and the solution to hospitals and also conducts mass-screening camps on its own. It charges a flat fee for the thermal camera and levies a user fee for the software. So far, it has sold its solution to 11 hospitals – in Bengaluru, Pune, Dehradun and Mysuru – and will shortly be selling it in Hyderabad and Chennai. Over 4,000 women have been screened in these hospitals so far.

Expansion plan

According to her, Niramai aims to grow both the diagnostic and screening businesses. The diagnostic business is when it sells the camera and the software to hospitals and the screening is what it does jointly with other organisations. Niramai has a 20-member team, which will grow to 40 in the next year or two.

Geetha says they will focus on India for the next two years and then try to sell the product to the global market. For this, Niramai has to get regulatory approvals for its products, for Europe and the US. It will look to raise more funds for international expansion. “Using Artificial Intelligence and Machine Learning, we are pretty much the only company doing it in a non-contact way. There are others trying to build wearables,” says Geetha, of Niramai’s strengths.