One of the key features of recently released Note 10 and Note 10+ is Samsung’s handwriting recognition solution

Figure 1 Handwriting into text feature in Note 10 Samsung Notes application

This year the solution has been presented at two international conferences – ICDAR in Sydney, Australia and UIST in New Orleans, USA – and has proven its status as the world-best handwriting recognition solution.

The 15th International Conference on Document Analysis and Recognition held on September 20-25, 2019 included CROHME + TFD Competition (Competition on Recognition of Handwritten Mathematical Expressions and Typeset Formula Detection) as a part of event. Samsung R&D Institute Ukraine has entered the competition with their handwritten mathematical expression recognition solution.

Figure 2 Handwritten formula recognition

Figure 3 The participation certificate
Recently, the organizers have shared official competition results. Samsung R&D Institute Ukraine was called a winner in TFD competition, had the highest score (First place) in formula parsing task, and took the second place (Runner up prize) in formula recognition task.

Figure 4 The results of  ICDAR2019 CROHME competition were announced in September 2019 in Sydney, Australia
Jury Statement:
“We summarize the tasks, protocol, and outcome for the 6th Competition on Recognition of Handwritten Mathematical Expressions (CROHME), which includes a new formula detection in document images task (+ TFD). More than thirty (33) participants registered for the competition, with nineteen (19) teams submitting results. System performance has improved since the last CROHME - still, the competition results suggest that recognition of handwritten formulae remains a difficult structural pattern recognition task.
For the new typeset formula detection task, the Samsung R&D Institute Ukraine obtained a very strong F-score (93%). There is a real gap between the winning system Samsung R&D and the other participants.”

ICDAR has become one of the most important international conferences in the community of pattern recognition and artificial intelligence. It covers the main themes of document analysis and recognition, handwriting analysis and verification, text detection and processing, and other related topics. Therefore this victory is very important point in field of Recognition development for Samsung Electronics.

In November 2019, two SRK’s papers on handwriting recognition technology have been presented at the 32nd ACM User Interface Software and Technology Symposium in New Orleans, USA. UIST is the premier world forum for innovations in human-computer interfaces. Samsung’s results were presented along with the work of such commercial and academic technology leaders as Microsoft, Google, Facebook, MIT, CMU, etc.

Figure 5 Discussion highlights at UIST2019, New Orleans, USA; ©ACM UIST2019

Figure 6 "I really enjoyed our discussion and seeing Samsung handwriting recognition and pen and touch demos"; © ACM UIST 2019
On reflection, the participants have found the experience beneficial both for the team and for Samsung. The competition served as an excellent ground for comparing Samsung's system to competitors' newest approaches (a rare opportunity with proprietary solutions). The results drew more attention to the team's work inside the company. The participation helped the team members better understand what it takes to present their system on a wider scale outside the company. The participation and state-of-the-art performance also established Samsung's reputation among other specialists in handwriting recognition.

Figure 7 SRK Handwriting Recognition Team
Presenting Samsung’s findings and approaches to wider scientific community is helping our researchers to broaden their horizons, in accordance to Samsung’s AI vision.