查詢結果分析
來源資料
相關文獻
- 線上手寫字辨認之模擬系統--利用類神經網路
- 多重專家應用於線上手寫中文文字識別
- 線上手寫中文文字辨認系統的筆段抽取
- On Speeding Radical-Based On-Line Chinese Character Recognition Through the Directional Matching of the First and Last Segments
- Character Recognition Using an Assembly of SSO Neural Networks
- An Overview of RNN-Based Mandarin Speech Recognition Approaches
- 以類神經網路為基礎之線上手寫中文字辨識系統
- 快速學習類神經網路手寫英數與中文字辨認系統
- 改良式漢明類神經網路應用於光學字元辨認
- An Artificial Neural Network System Process Control
頁籤選單縮合
題 名 | 線上手寫字辨認之模擬系統--利用類神經網路 |
---|---|
作 者 | 鄭錦秋; | 書刊名 | 南臺工商專校學報 |
卷 期 | 17 1993.03[民82.03] |
頁 次 | 頁45-62 |
分類號 | 312.74 |
關鍵詞 | 手寫字; 原始程式碼; 模擬系統; 線上; 辨認; 類神經網路; |
語 文 | 中文(Chinese) |
中文摘要 | 本文旨在利用類神經網路之強韌的容錯度,建立一套可辨認任何手寫字型之模擬 系統。 採用之類神經網路類似於 K. Fukushima 的 Neocognitron,在網路學習方面,神經 鍵只記錄是否有被激發,而不考慮計算神經鍵值之數值大小,如此,可加速學習之速率。在 網路辨認方面,考慮前景與背景之關係,以增強不同類字形間之差異;另外亦考慮局部濃度 分布,以增強同類字型之模糊容忍度。 經模擬系統測試結果顯示,本系統在學習速率方面,相當快速,而在辨認方面,有學習過之 字型,辨認成功率 100 %, 未學習過之字型, 只要筆跡端正不潦草,辨認成功率可達 98 %以上。惟在辨認速率方面,稍嫌遜色,是尚待改進之處。 |
英文摘要 | Thr purpose of the paper is to develop a simulation system of online handwritten character recognition by the though tolerance of architecture neural network. In the paper, the adapted neural network is similar to the Neocognitron of K. Fukushima. In the network learning, the neural weights are recorded only and not calculated when the related neurons are excited. Thus, it will advance the speed of the network learning. In the network recognition, we consider the interrelation between the front and the rear to enhance the of disparate characters. In addition, we will consider the distribution of local concentration in order to enhance the fuzzy tolerance of akin characters. By testing the simulation system, the results tell us that the learning speed of the system is very fast and the recognition rate is up to 100%, For the characters which are not yet learned by the system, the recognition rate could be 98% when the fist of the character is righteous. But the recognition speed is poor and need to improve. |
本系統中英文摘要資訊取自各篇刊載內容。