The researchers at University College London have designed a smart algorithm able to analyze handwritten text and to clone that style.
Handwriting Replication Algorithm
Handwriting recognition programs have been developed before, one example is Samsung's built-in handwriting recognition software that allows the Note 3 to transform users' handwriting into text using the S Pen. However, the new smart application takes the idea further and makes it possible to actually clone handwriting.
According to Gizmodo, the new algorithm designed by the researchers at University College London is built around a specific instance of a character called glyph. The research team has published their study on the university's website.
Digital Trends reports that the new machine learning tool has been called "My Text In Your Handwriting" by the UCL researchers who designed it. The handwriting replicator algorithm works by first marking up each punctuation mark and letter.
The smart app is able to recognize a person's specific character vertical, choices, spacing, texture and the inter-character ligatures. The software chooses the average, since every person's handwriting has slight differences within itself.
Potential Applications Of The Handwriting Replicator Algorithm
The lead author on the project, Tom Haines, said in a statement that the software developed by its research team has the potential to help banks or other institutions that want to send out sensitive documents that look like handwritten letters, or people who have trouble holding a pen.
The handwriting replicator software could also be used in comic books where handwritten text could be automatically translated into other languages. The smart application can also have potential benefits to better register signatures of those who need to sign contracts online.Study author Dr. Gabriel Brostow noted that, since the writing can be analyzed under a microscope, the system can also be used to detect forgeries. He explained that the application can be used to characterize handwriting in order to evaluate the odds that something was forged.
