Announcement for Downloading full text filePlease respect the Copyright Act.
All digital full text dissertation and theses from this website are authorized the copyright owners. These copyrighted full-text dissertation and theses can be only used for academic, research and non-commercial purposes. Users of this website can search, read, and print for personal usage. In respect of the Copyright Act of the Republic of China, please do not reproduce, distribute, change, or edit the content of these dissertations and theses without any permission. Please do not create any work based upon a pre-existing work by reproduction, Adaptation, Distribution or other means.
URN etd-0830104-185208 Statistics This thesis had been viewed 3266 times. Download 2888 times. Author Wei-Chih Pan Author's Email Address firstname.lastname@example.org Department Computer Science and Enginerring Year 2003 Semester 2 Degree Master Type of Document Master's Thesis Language English Page Count 41 Title Application of the Region Growing Algorithm on Handwritten Numeral Recognition Keyword Region Growing Handwritten Numeral Recognition Handwritten Numeral Recognition Region Growing Abstract In the handwritten numeral recognition system, researchers use a lot of features to help the recognition process, such as end points, fork points, strokes, circles, etc. These features were extracted from the original image as the basis for the numeral recognition. In this thesis, we propose a handwritten numeral recognition algorithm that based on the Region Growing algorithm, which is often applied in the image retrieval researches. With proper modification, this algorithm can also be applied to the handwritten numeral recognition.
The Region Growing algorithm can simplify the image information based on the attribute of pixels of the original. Adjacent pixels with similar gray scale can be combined into the same region. After this operation, there will be spatially separable regions. Then based on the spatial distribution of these regions, we can find the similarity between images. Finally, we can classify and recognize the image according to the region’s distribution and similarity.
Furthermore, we proposed a modified “Drop Falling algorithm” to deal with the segmentation problem of the original image. This algorithm in conjunction with the histogram can make proper cut trace and get accurate recognition results
The experimental results show our proposed algorithm can achieve the accuracy requirement of handwritten numeric recognition system.
Advisor Committee Tsang-Long Pao - advisor
Shang-Lin Hsieh - co-chair
Files Date of Defense 2004-07-26 Date of Submission 2004-08-30