DeepQuanty’s AI-powered engine can read handwriting on cheques

Chennai-based DeepQuanty Artificial Intelligence Labs has announced the launch of its first product, an artificial intelligence-powered engine that can decode text on cheques.

The recently-floated firm said in a statement that the product, SnapChek, can deliver 90% extraction accuracy on any handwritten fields in cheques such as date, name, amount in words and amount in figures, and near 100% extraction accuracy in printed fields such as account number, MICR and IFSC Code.

“SnapChek provides banks in India a significant opportunity to enhance banking experience,” said Mahadevan Jayaram, co-founder & chief revenue officer of DeepQuanty.

SnapChek was developed at DeepQuanty’s India development centre by a team of AI specialists in under four months, the statement added.

“Our AI team is proud to have solved a very old handwriting recognition problem in an unconventional fashion, which puts SnapChek way ahead in performance when compared to known and published benchmarks,” said Sundar Kanniappan, chief technology officer of DeepQuanty.

The startup said that it would launch more products in the coming months.

DeepQuanty Artificial Intelligence Labs was set up four months ago by Kanniappan and Deepak Ramesh with the objective of applying computer vision and artificial intelligence to solve problems in the banking, retail, e-commerce, insurance and healthcare sectors.

Their aim was to extract data from images such as cheques, application forms, identity proof, and apparel, with an ability to build further intelligence post extraction.