糖心视频

One mislabeled image can mean difficulty in finding it later, something is working to change so that data scientists can find what they are looking for the first time.

Subscribe to the Crunchbase Daily

The San Francisco-based company is developing an artificial intelligence-powered annotation platform for data scientists and labeling teams. On Thursday the company announced it had raised $3 million in a seed funding round led by , with participation from and . Previous investors and also participated.

The new funding gives the 2-year-old company $3.7 million in total funding raised, which includes $700,000 in pre-seed and various other stages of early investment, CEO told Crunchbase News.

Petrosyan, who has a background in biomedical engineering, and his brother, , CTO, built SuperAnnotate on technical algorithms Vahan was developing as part of his Ph.D. work.

The market for data annotation tools is expected to reach, according to Meanwhile, a March 2019 report by found that data preparation and engineering tasks performed in most AI and machine-learning projects represented.

The company鈥檚 end-to-end platform for image annotation enables users to manage the entire process of manual labeling, including a toolset for handling communications, responsibilities and workflows. The platform went live in February and has already attracted more than 3,000 data scientists and over 100 companies, Petrosyan said.

Petrosyan explained that labeled data is important to advance industries such as self-driving cars, retail and health care, so that certain images and products are recognized, and in the case of health care, images are ruled out as cancerous.

鈥淚maging annotation is a problem because most organizations have tens of thousands to millions of images that need to be labeled,鈥� Petrosyan said. 鈥淣ot just one person does it, but you could have tens to thousands put in charge of labeling, not only as fast as possible, but accurately. If an image is labeled incorrectly, it will be bad.鈥�

The new round of funding will be used to expand SuperAnnotate鈥檚 engineering, sales and marketing teams. The company has 38 full-time employees, and Petrosyan would like to see it hit about 50 by the end of the year. On the tech side, the new funds will go toward building out more tools for video annotation, 3D tools and machine-learning features.

Next up, the company is scaling to serve large companies, including those that need computer vision and integrated machine-learning features.

鈥淭hough we launched three months ago, we have been overwhelmed by the organic adoptions of our product, and we are ready to serve many industries that need computer vision and also integrated machine-learning features,鈥� Petrosyan said.

Illustration: