登入選單

返回Google圖書搜尋

Image Thresholding of Historical Documents Based on Genetic Algorithms

Carmelo Bastos FilhoAdriano OliveiraCarlos Alexandre MelloDavi FalcaoJulio AndradeMarilia LimaWellington dos Santos

出版	INTECH Open Access Publisher, 2008
ISBN	95376190369789537619039
URL	http://books.google.com.hk/books?id=3_S1oAEACAAJ&hl=&source=gbs_api

註釋

It is presented in this paper a technique for automatic thresholding images of historical documents, in special, documents written on both sides of the paper, presenting back-tofront interference. The new method uses genetic algorithms to achieve a quantized image and proceed with a binarization. The resulting images were analyzed using a fidelity index, PSNR and measures from signal detection theory. The method can also be extended to optimize quantization processes for other types of images. The method proved to be more efficient than several other classical thresholding algorithms in an evaluation using precision, recall, accuracy and specificity. Currently the bi-level images are being used to improve several steps on an optical character recognition process of these documents such as text segmentation and the recognition per se.