摘要
提出一种基于小波重构的视频图像文本提取方法.原始图像经小波分解重构得到特征图像,在特征图像上使用固定尺寸的文本检测器提取图像区域的属性特征,利用神经网络分类器确定该图像区域的属性.实验表明,本方法与基于边缘特征的文本提取方法相比,具有更好的性能.
A novel method for text detection in image and video based on wavelet reconstruction is proposed. Feature image is reconstructed by wavelet decomposition of the original image, in which text region property are confirmed by using neural network trained to extract property feature by a fixed size text detector, compared with text detection method based on edge features, the proposed method has a better accuracy.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2006年第9期1458-1460,共3页
Journal of Harbin Institute of Technology
关键词
文本检测
文本定位
小波重构
神经网络
Text detection
Text location
Wavelet reconstruction
Neural Network
