Text Localization Using Discrete Wavelet Transform and Neural Network

• Main Authors: Chung-Wei Liang, 梁忠瑋
• Other Authors: Po-Yueh Chen
• Format: Others
• Language: en_US
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/68u4za

碩士 === 朝陽科技大學 === 資訊工程系碩士班 === 92 === Texts provide highly condensed information about the contents of images or video sequences. Although texts provide important information about images and video sequences, it is not an easy problem to detect and segment them. Most of the text localization methods were applied to uncompressed images. Only a few of them proposed to locate texts from the compressed images. In this thesis, we propose a text localization method using discrete wavelet transform and neural network. Because the DWT coefficients provide important information of text regions, we employ those coefficients and several image processing techniques to preliminarily locate candidate text regions. Then the morphological dilation operation and the neural network are employed to raise the recall rate and precision rate. According to the experimental results, the proposed method can successfully locate text region from complex images. Its recall rate and precision rate are better than that of other methods.