发明名称 |
Text detection in natural images |
摘要 |
A system and method of text detection in an image are described. A component detection module applies a filter having a stroke width constraint and a stroke color constraint to an image to identify text stroke pixels in the image and to generate both a first map based on the stroke width constraint and a second map based on the stroke color constraint. A component filtering module has a first classifier and second classifier. The first classifier is applied to both the first map and the second map to generate a third map identifying a component of a text in the image. The second classifier is applied to the third map to generate a fourth map identifying a text line of the text in the image. A text region locator module thresholds the fourth map to identify text regions in the image. |
申请公布号 |
US9076056(B2) |
申请公布日期 |
2015.07.07 |
申请号 |
US201313970993 |
申请日期 |
2013.08.20 |
申请人 |
ADOBE SYSTEMS INCORPORATED |
发明人 |
Wang Jue;Lin Zhe;Yang Jianchao;Huang Weilin |
分类号 |
G06K9/18 |
主分类号 |
G06K9/18 |
代理机构 |
Shook, Hardy & Bacon L.L.P. |
代理人 |
Shook, Hardy & Bacon L.L.P. |
主权项 |
1. A method comprising:
applying a filter having a stroke width constraint and a stroke color constraint to an image to generate a first map based on the stroke width constraint and a second map based on the stroke color constraint; applying a first classifier to both the first map and the second map to generate a third map identifying a component of a text in the image; applying a second classifier to the third map to generate a fourth map identifying a text line of the text in the image, wherein the second classifier comprises a Text Covariance Descriptor (TCD) algorithm for text lines, the TCD algorithm configured to:
compute a first covariance matrix for correlated features of components between heuristic and geometric properties of the first map and the second map,compute a second covariance matrix for correlation of statistical features among the components,generate a feature vector based on the first and second covariance matrices, a normalized number of components in a text-line, and a mean confidence score of the third map, andgenerate a confidence score of the fourth map for each text-line candidate from the feature vector; and thresholding the fourth map to identify text regions in the image. |
地址 |
San Jose CA US |