空間拓撲距離測量和形狀規則手寫字符識別

Spatial Topology Distance Measurement and Shape Rules for Handwritten Character Recognition

葉金龍(KIM-LONG DIEP)

指導教授 ： 劉長遠

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2016年)

https://doi.org/10.6342/NTU201600391

摘要

這項工作表明如何設置形狀規則，並將其轉換為邏輯規則，跳過不正確的模板和減少空間拓撲變形方法[1,2,3,4,5]的候選模板的數量。識別率也由包括在自組織過程形狀限制改善。這將極大地降低具有改善識別計算的數量。
我們複習了一種自機構匹配方法來完成的厚筆觸繪製手寫字符識別。這種方法被用來針對彎曲逐漸走向，匹配其對象中的字符未知字符手寫。
被提取的字符的特徵被使用在自機構匹配是中心軌跡，方向，和橢圓的主軸, 他適合的圖案的著墨面積。提供了使用建議方法得到模擬值得獎勵結果。

關鍵字

手寫字符識別,空間拓撲距離,自機構映射,神經網絡,彈性匹配,模式識別,自組織地圖,變形處理,形狀規則,形狀限制.

並列摘要

This work shows how to set shape rules and convert them into logical rules to skip incorrect templates and reduce the number of candidate templates in the spatial topology distortion method [1, 2, 3, 4, 5]. The recognition rate is also improved by including shape constraints in the self-organizing process. This will drastically reduce the number of computations with improved recognition.
We study a self-organization matching approach to accomplish the recognition of hand-printed characters drawn with thick strokes. This approach is used to flex the unknown hand-printed character toward matching its object characters gradually. The extracted character features used in the self-organization matching are center loci, orientation, and major axes of ellipses which fit the inked area of the patterns. Simulations provide encouraging results using the proposed method.
In the spatial topology distortion method [1], named STD, the distortion between a candidate template and an unknown pattern can be computed by using the self-organizing algorithm [7]. This distortion is used to rank its candidate. It is cost to obtain such fine distortions for all templates. It is expected that certain incorrect templates can be skipped by imposing the rules among template features. The STD will be operated only for those templates that meet the rules for fine discrimination.
We briefly review the hand-printed character recognition techniques for thick strokes and discuss their difficulties. The difficulties are mainly arisen from the various flexible distortions produced during handwriting. Robust techniques on the thinning method, correlation matching, elastic matching, and distance measurement are the main focuses for solving such difficulties.

並列關鍵字

Hand-printed character recognition, spatial topology distance, self-organizing map, neural networks, elastic matching, pattern recognition, self-organizing map, morphing process, shape rule, shape constraint.

參考文獻

[5]. Daw-Ran Liou, Chia-Ching Lin and Cheng-Yuan Liou (2012), Setting Shape Rules for Handprinted Character Recognition, ACIIDS, The 4rd Asian Conference on Intelligent Information and Database Systems, March 19-21, LNCS 7197, pp. 245-252, Kaohsiung, Taiwan.

[1]. D.J. Burr (1981). “Elastic Matching of Line Drawings”, IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 3, pp. 708-713.

[2]. Cheng-Yuan Liou and Hsin-Chang Yang (1999). Selective feature-to-feature adhesion for recognition of cursive handprinted characters, IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 21, no. 2, pages 184-191. SCI&EI.

[3]. Cheng-Yuan Liou and Hsin-Chang Yang (1996). Handprinted character recognition based on spatial topology distance measurement. IEEE Trans. on Pattern Analysis and Machine Intelligence, vol.18, no.9, pages 941-945. SCI&EI.

[6]. D.J. Burr (1983), “Designing a Handwriting Reader, ” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 5, pp. 554-559.