Character correction method and device, terminal equipment and storage medium

Abstract

The invention discloses a character correction method and device, terminal equipment and a storage medium, and the method comprises the steps: obtaining writing data, and converting the writing data into a stroke sequence corresponding to an input character; segmenting the stroke sequence according to inflection points to obtain sub-stroke sequences corresponding to sub-strokes of the input character; correcting the sub-strokes based on the sub-stroke sequence to obtain a corrected sub-stroke sequence of the sub-strokes; and splicing the corrected sub-stroke sequences to obtain the corrected stroke sequence of the input character. According to the scheme, correction is carried out based on the finer-grained sub-stroke level, the final character recognition result is improved, the interaction experience is enhanced, and the working overhead of the correction scheme can be greatly reduced.

Description

technical field [0001] The present application relates to the technical field of information processing and human-computer interaction, and in particular to a character correction method, device, terminal equipment and storage medium. Background technique [0002] Artificial intelligence technology is gradually applied in all walks of life. The combination of artificial intelligence and human-computer interaction has attracted a lot of research in recent years, one of which is hot: using inertial sensors for handwriting in the air. [0003] Aerial handwriting is a technology that directs hands and arms, and converts and recognizes hand movement trajectories into corresponding information, which can be realized by wearing or holding devices equipped with inertial sensors. Different from the way of writing on paper and touch screens of electronic devices in daily life, the movement range of waving the hand is much larger than that of writing with the tip of the pen. Therefore...

AI Technical Summary

Problems solved by technology

These two types of writing noise will affect the character recognition results, and non-standard glyphs will also affect the interactive experience

[0004] When the related technology corrects the handwritten characters in the air, it usually corrects the overall tilt or jitter of the characters, resulting in poor character recognition results, and it is necessary to obtain a large amount of data in advance for statistics or modeling learning, and the amount of offline calculation is large, resulting in a correction scheme. high overhead

Images