Text detecting method and apparatus

Abstract

The invention provides a method for calculating the estimated stroke width (ESW). The method comprises the steps of obtaining stroke edge information according to a binary image; calculating the stroke width of each stroke edge pixel point in not less than four orientations, with the stroke width of each stroke edge pixel point in the not less than four orientations being the distance from the stroke edge pixel point to another stroke edge pixel point located on a straight line determined by the stroke edge pixel point and the orientations; associating the stroke width of each stroke edge pixel point in the not less than four orientations with each stroke interior pixel point passing the stroke edge pixel point and along the orientations; and for each stroke interior pixel point, selecting the minimum value of the plurality of stroke width values associated with the stroke interior pixel point as the ESW of the stroke interior pixel point. The invention further provides an apparatus corresponding to the method.

Description

technical field [0001] The invention relates to human-computer interaction technology, in particular to text detection or optical character recognition (OCR) technology. Background technique [0002] Natural scenes not only contain a large amount of graphic information, but also rich text information, such as road signs, store names, etc. These text information are of great value to the description and understanding of scene content, and this information is the key clue of scene image retrieval. Therefore, an automated tool is urgently needed to obtain text information in the scene through text recognition in natural scenes, serve for retrieval, query, browsing scene image data and understand scene content, and improve the management efficiency of image data. Mobile phones, PDAs, desktop computers, laptops, tablets, and other electronic devices often support text detection or optical character recognition (OCR). [0003] Stroke Width Transform (SWT) is a commonly used text...

AI Technical Summary

Problems solved by technology

[0005] However, by analyzing the above SWT algorithm, it is easy to know that the algorithm has the following problems: because the stroke edge is in an irregular shape, it is a very complicated process to calculate the tangent direction at the pixel point p in step S120. The complexity is high and consumes a large amount of processor resources and calculation time; in step S150, the stroke width w is compared with the value a stored in association with the pixel point t, however, due to the large number of stroke edge points and the irregular shape of the stroke edge, the points inside the stroke There may be multiple normals passing by, which will cause too many comparisons and the processing is very cumbersome

[0006] Therefore, the text detection process using SWT is too complicated and time-consuming. The prior art mentions that the text detection time is 0.94 seconds, and for the OCR system of natural scenes, the OCR processing process after text detection also takes time , and there are further applications in the future, such as translation or retrieval, etc., so the speed of SWT is too slow as a preprocessing step in the OCR system, and it is far from meeting the real-time requirements of the natural scene OCR system

Images