New Computer Programme Replicates Handwriting

Software – `My Text in Your Handwriting’ Programme

In the present world which is controlled by the QWERTY keyboard. Computer scientists of UCL have come up with software which may ignite the return of the handwritten word by studying the handwriting of any person and tend to imitate it precisely. The scientist have generated `My Text in Your Handwriting’, a programme that is said to semi-automatically test a sample of handwriting of a person which could be a small paragraph, generating new text stating that whatever the user desires, would be as though the author had handwritten it. Dr Tom Haines of UCL Computer Science had stated that the software had lots of valued application.

 For instance those suffering from strokemay be capable of formulating letters with the need of illegibility or the one sending flowers as a token of gift could send a handwritten note without even going to the florist. Besides this, it could also be utilised in comic books wherein a part of handwritten content could be translated into various languages without the need of losing the original style of the author. The machine learning process, published in ACM Transactions on Graphic and funded by the EPSRC, has been built around glyphs, which is an explicit example of a character.

Authors Create Various Glyphs – Element of Writing

Authors tend to create various glyphs in representing the same element of writing, the way one individual tends to write an `a’ would generally differ with the way others tend to write an `a’. Though the writing of an individual tends to have some variations, each author seems to have a recognisable style which demonstrates in their glyphs together with their spacing.

The software absorbs what is stable through the style of the individual and reproduces it. In order to produce the handwriting of an individual, the program tends to analyse and duplicate the specific character choices of the author, its pen-line texture as well as colour, inter-character ligatures, the joining between letters, the vertical and horizontal spacing. Dr Oisin Mac Aodha, co-author of UCL Computer Science had stated that till now the only way to create a computer generated text which is similar to a particular person’s handwriting could be by utilising an applicable font.

System Adequately Flexible

The issue with these types of fonts is that it frequently tends to clear that the text is not penned by hand and that loses the character as well as the personal touch of a handwriting part of the text. What has been invented is what removes this issue and could be utilised in a wider variety of personal and commercial situations.

The system is said to be adequately flexible where the samples from historical documents could be utilised with the minimum effort and hence the scientists have scrutinized and simulated the handwriting of figures like Abraham Lincoln, Arthur Conan Doyle and Frida Kahlo. In order to assess the efficiency of the software, the team had requested people to differentiate between handwritten envelopes and the ones created by their automatic software where the people were baffled by several of the envelopes.

This resulted in being fooled by the computer generated writing by 40% of the time. Taking into consideration how conclusive the computer generated handwriting seems to be, some are of the belief that the method could be beneficial in forging documents, though the team clarified that it seems to work both way and could be essentially supportive in distinguishing forgeries.