公开(公告)号：US09367736B1

公开(公告)日：2016-06-14

申请号：US14842125

申请日：2015-09-01

Applicant: Amazon Technologies, Inc.

Inventor： Thibaud Senechal ,  Quan Wang ,  Daniel Makoto Willenson ,  Shuang Wu ,  Yue Liu ,  Shiv Naga Prasad Vitaladevuni ,  David Paul Ramos ,  Qingfeng Yu

IPC: G06K9/46 ,  G06K9/00 ,  G06K9/34

CPC classification number: G06K9/00463 ,  G06K9/00442 ,  G06K9/00456 ,  G06K9/344 ,  G06K9/348 ,  G06K9/4638 ,  G06K9/4652 ,  G06K2209/01

Abstract: A multi-orientation text detection method and associated system is disclosed that utilizes orientation-variant glyph features to determine a text line in an image regardless of an orientation of the text line. Glyph features are determined for each glyph in an image with respect to a neighboring glyph. The glyph features are provided to a learned classifier that outputs a glyph pair score for each neighboring glyph pair. Each glyph pair score indicates a likelihood that the corresponding pair of neighboring glyphs form part of a same text line. The glyph pair scores are used to identify candidate text lines, which are then ranked to select a final set of text lines in the image.

Abstract translation: 公开了一种多方向文本检测方法和相关系统，其利用取向变体字形特征来确定图像中的文本行，而不管文本行的取向如何。 为相对于相邻字形的图像中的每个字形确定字形特征。 字形特征被提供给学习的分类器，其为每个相邻字形对输出字形对分数。 每个字形对得分表示对应的相邻字形对形成相同文本行的一部分的可能性。 字形对分数用于识别候选文本行，然后将其排序以选择图像中的最后一组文本行。