Esoft submits first ever scientific paper

As part of his Industrial Ph.D. at Esoft, Juan Francisco Marin Vega has submitted a paper to one of the leading computer vision conferences in the world. This also marks an important milestone for Esoft’s AI ambitions.

Earlier this year, Esoft as a company reached an important milestone in our transition from IT company to tech company.

In November, Industrial Ph.D candidate Juan Francisco Marin Vega, from Esoft’s data science department, submitted a scientific paper to the CVPR 2021 conference – one of the most important computer vision and pattern recognition conferences in the world.

The list of approved papers for CVPR21 won’t be made public until January 2021. Regardless of the outcome, Esoft CIO Michael Sloth calls the submission a very important milestone for the company.

It goes to show that when we talk about becoming a tech company, we mean it – seriously,” says Sloth. “We aim to bring AI into real estate, and we do that by introducing world-class R&D into our daily operations.”

High refusal rate

For Marin Vega personally, the submission of the paper is similarly a big deal. “It is also the first time that I’ve submit a paper like this, so I’m very happy that we are now at a stage where our work is actually being reviewed by some very skilled people,” he says.

Marin Vega adds that the refusal rate for submitted papers is really high, around 80%. For example, for the 2019 conference, a total of 6,700 papers were submitted but only 1,500 were approved.

Merging science with business

The paper deals with one of Marin Vega’s major focus areas in the first year of his Ph.D. – single exposure image enhancement. This work also serves as the theoretic framework for the AI-based photo editor that Esoft recently introduced into production.

“What we have submitted in the paper is just the science part of it. But it is a good example of how we try to incorporate science into our core business processes in order to improve efficiency and quality,” Marin Vega concludes.

 

 

Next story

Chasing the perfect image: Esoft launches its first AI solution