Priority Of The Chinese Forum Of Non-text-based Image Analysis Technology Research And Application

OCR recognition rates up to 100 percent for general layout of text. As for the photo-text layout, OCR recognition need layout analysis, isolated text, tables, graphics, images and even the title, formula, columns and other layout elements and dealt with accordingly.The Chinese needed to solve the layout of the difficulties than Spanish as much as the style of Chinese publications, more and more pages are constantly changing, layout within the text, images, interlacing, horizontal, vertical version of fusion, each block of text font, size also frequent changes in the interval between the line of text block, lace line range. Again because the characteristics of Chinese characters themselves, the majority of characters consists of several parts of the radical form, its easy and punctuation with the radical and confusing on the page segmentation and region identification region caused great disturbance, giving the Chinese page layout analysis and layout also add the difficulties identifiedLayout analysis as a prerequisite for document recognition and recovery, the fitness of Layout analysis algorithms directly affects the results of document recognition and recovery.This thesis are being used in Blind Readers which was developed by "Eleventh Five-Year" National Science and Technology Support Program Key Project ". The layout analysis as an important aspect of layout analysis algorithms required to solve complex layout of the situation and the algorithm efficiency and algorithm accuracy have certain requirements.This thesis studies various layout analysis algorithms and summaries their advantages and disadvantages on applying to Blind Reader. The thesis improve the existing layout analysis algorithm, using recursive projection method assisted connected region approach, making division of the same layout, the image separation of regional priorities and address them through the connected components of the text area to optimize the projection method to solve the projection method can not handle the general situation of Chinese layout, connected region approach slowly.