本文選題：公式識(shí)別 + SVM�。� 參考：《沈陽(yáng)工業(yè)大學(xué)》2015年碩士論文

【摘要】：光學(xué)字符識(shí)別（OCR）是近年來(lái)廣泛應(yīng)用于銀行、郵電、物流等領(lǐng)域的一種識(shí)別技術(shù)，目的是將以圖像方式輸入的印刷體或手寫體字符轉(zhuǎn)化為可編輯的符號(hào)。目前，對(duì)印刷體文檔中的中英文以及阿拉伯?dāng)?shù)字的識(shí)別已達(dá)到較高水平，但由于數(shù)學(xué)公式符號(hào)的種類多，變化大，結(jié)構(gòu)復(fù)雜，實(shí)現(xiàn)正確、快速的識(shí)別比較困難，需要探索更有效的識(shí)別方法。 本文針對(duì)印刷體數(shù)學(xué)公式識(shí)別中的幾個(gè)關(guān)鍵問(wèn)題展開(kāi)研究，重點(diǎn)解決傾斜圖像的快速與準(zhǔn)確校正、粘連符號(hào)的有效分割和基于SVM的多層分類器的符號(hào)識(shí)別問(wèn)題。為了提高版面傾角檢測(cè)的效率和精度，提出了一種基于連通域分析和Hough變換的傾斜校正方法，通過(guò)連通域分析預(yù)估傾角，以較長(zhǎng)的連通域?yàn)橐罁?jù)劃分出文本區(qū)域，結(jié)合經(jīng)邊緣檢測(cè)處理后的版面區(qū)域，以不同角度步長(zhǎng)分別進(jìn)行Hough變換，得到最終精確的傾角。同時(shí)，通過(guò)凹凸輪廓和分割因子確定待分割位置，進(jìn)而對(duì)分割后的符號(hào)進(jìn)行識(shí)別驗(yàn)證。由于公式符號(hào)眾多，為了有效地降低分類器的負(fù)擔(dān)并提高分類的準(zhǔn)確性，對(duì)公式符號(hào)的特征進(jìn)行詳細(xì)篩選和分類，并以此為基礎(chǔ)構(gòu)造了粗、細(xì)分類相結(jié)合多層分類器。在細(xì)分類時(shí)，利用一對(duì)多的方法改進(jìn)了傳統(tǒng)DAG-SVM訓(xùn)練模型中的一對(duì)一方法，提高了分類器的訓(xùn)練效率，并利用類間可分性對(duì)DAG-SVM中的節(jié)點(diǎn)進(jìn)行重新組合，降低了誤差累積對(duì)識(shí)別所造成的影響。實(shí)驗(yàn)和分析表明，所提出的算法能夠高效檢測(cè)出版面的傾角，實(shí)現(xiàn)準(zhǔn)確的粘連字符分割，完成有效的公式符號(hào)識(shí)別。 基于上述方法，，本文應(yīng)用VC++設(shè)計(jì)并實(shí)現(xiàn)了一個(gè)印刷體數(shù)學(xué)公式識(shí)別系統(tǒng)。以包含公式的文檔圖像作為系統(tǒng)的輸入，經(jīng)過(guò)版面分析、公式圖像預(yù)處理、公式符號(hào)識(shí)別和公式結(jié)構(gòu)分析，將其以Latex的格式輸出。通過(guò)對(duì)識(shí)別結(jié)果的分析，使用本文所提出的改進(jìn)的SVM分類器對(duì)數(shù)學(xué)符號(hào)進(jìn)行識(shí)別，可以達(dá)到94.7%的識(shí)別率，要高于現(xiàn)有的SVM分類器的識(shí)別率。
[Abstract]:Optical character recognition (OCR) is a recognition technology widely used in the fields of bank, post and telecommunication, logistics and so on in recent years. The aim is to convert printed or handwritten characters input by image into editable symbols. At present, the recognition of Chinese and English and Arabic numerals in printed documents has reached a high level, but due to the large variety of mathematical formula symbols, large changes, complex structure, it is difficult to realize correct and rapid recognition. More effective identification methods need to be explored. In this paper, several key problems in the recognition of printed mathematical formulas are studied, focusing on the problems of fast and accurate correction of skew images, effective segmentation of adhesive symbols and symbol recognition of multi-layer classifiers based on SVM. In order to improve the efficiency and accuracy of layout dip detection, a tilt correction method based on connected domain analysis and Hough transform is proposed. Combined with the area of layout after edge detection, the Hough transform is carried out with different angle step sizes, and the final inclination angle is obtained. At the same time, the position to be segmented is determined by the concave and convex contour and the segmentation factor, and then the symbol after segmentation is recognized and verified. In order to effectively reduce the burden of classifiers and improve the accuracy of classification, the features of formula symbols are screened and classified in detail. Based on this, a coarse and fine classifier combined with multi-layer classifier is constructed. In fine classification, the one-to-one method of traditional DAG-SVM training model is improved by one-to-many method, and the training efficiency of classifier is improved, and the nodes in DAG-SVM are recombined by using inter-class separability. The effect of error accumulation on recognition is reduced. Experiments and analysis show that the proposed algorithm can efficiently detect the dip angle of the layout, achieve accurate segmentation of the adherent characters, and achieve effective formula symbol recognition. Based on the above methods, a printing mathematical formula recognition system is designed and implemented by VC. The document image containing the formula is taken as the input of the system. After layout analysis, formula image preprocessing, formula symbol recognition and formula structure analysis, it is output in the format of Latex. Based on the analysis of the recognition results, the improved SVM classifier proposed in this paper is used to recognize the mathematical symbols. The recognition rate of the improved SVM classifier can reach 94.7%, which is higher than that of the existing SVM classifier.
【學(xué)位授予單位】：沈陽(yáng)工業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2015
【分類號(hào)】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 吳一全;謝靜;;基于特征點(diǎn)最小距離擬合的文檔圖像傾斜檢測(cè)[J];光學(xué)技術(shù);2009年01期

2 徐曉蓉;陳洪波;王強(qiáng);李永軍;;數(shù)學(xué)公式中數(shù)學(xué)符號(hào)的特征分析及提取[J];蘭州理工大學(xué)學(xué)報(bào);2012年05期

3 安艷輝;董五洲;;基于識(shí)別反饋的粘連字符切分方法研究[J];河北省科學(xué)院學(xué)報(bào);2008年02期

4 田學(xué)東;張艷;王亮;;基于輪廓特征的印刷公式粘連符號(hào)的切分[J];計(jì)算機(jī)工程與應(yīng)用;2008年04期

5 張順利;李衛(wèi)斌;吉軍;;基于投影的文檔圖像傾斜校正方法[J];計(jì)算機(jī)工程與應(yīng)用;2010年03期

6 計(jì)春雷;馮偉;黎明;楊杰;;一種動(dòng)態(tài)閾值加填補(bǔ)的指紋圖像二值化算法[J];計(jì)算機(jī)仿真;2011年07期

7 江明;劉輝;黃歡;;圖像二值化技術(shù)的研究[J];軟件導(dǎo)刊;2009年04期

8 李剛;宋文靜;;基于圖像直方圖的車牌圖像二值化方法研究[J];交通運(yùn)輸系統(tǒng)工程與信息;2009年01期

9 馮星奎,李林艷,顏?zhàn)嫒?一種新的指紋圖象細(xì)化算法[J];中國(guó)圖象圖形學(xué)報(bào);1999年10期

10 瞿洋,楊利平;Hough變換OCR圖象傾斜矯正方法[J];中國(guó)圖象圖形學(xué)報(bào);2001年02期

本文編號(hào)：1824384