Compressive Description of Digitized Voice or Music Data by means of Cellular Automata

通过元胞自动机对数字化语音或音乐数据进行压缩描述

• 批准号: 11837012
• 负责人: NARA Shigetoshi
• 金额: $ 2.3万
• 依托单位: Okayama University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 1999
• 资助国家: 日本
• 起止时间: 1999 至 2000
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-11837012/
• 关键词: セル・オートマトン; ルールダイナミックス; ビットパターン; 情報圧縮; 音声・音楽・動画像; 複雑さ; 圧縮率; 信号対雑音比; 音声・音楽

Under the title of the project, we have made various simulations and the obtained results were reported in the several conferences, workshops and also the several manuscripts were submitted to a few academic journals. A few papers were already published. The most remarkable point in our results is to have found the fact that only two rules of the two states and three neighbors cellular automata (referred as 1-2-3 CA hereafter) are sufficient to realize the error-less description of the digitized data which is coded by the binary format with the sampling frequency 44.1 kHz and with the 16-bit quantization, which are both the standard format of compact disc (CD) in the modern electronic technology. In the other words, this means that, once an initial bit pattern is given, then the whole data are completely reproduced by sequential applying of the rules of the 1-2-3 CA to the initial bit pattern. Furthermore the total amount of data necessary to describe the given data are less than the original data. Therefore, the data compression without loss of information is possible with our method, which was proposed as a novel idea created in this project. In order to investigate this topic in more detail, we have held a small workshop in November, 2000, where the place held the workshop is located near Okayama University. The invited speakers were(1) Masato Wada ; Hiroshima University(2) Takayuki Arima ; Okayama University(3) Song-Ju Kim ; Waseda University(4) Yoshinori Nagai ; Waseda University(5) Yoji Aizawa ; Waseda UniversityThe other participants are Jousuke Kuroiwa (Hiroshima University), Shigetoshi Nara (Okayama University) and the several graduated students in Okayama University.

在该项目的名义下，我们进行了各种模拟，得到的结果在多个会议、研讨会上进行了报告，并向一些学术期刊提交了多篇手稿。一些论文已经发表。我们的结果中最引人注目的一点是发现仅两个状态和三个邻居元胞自动机的两个规则（以下简称1-2-3 CA）就足以实现数字化数据的无错误描述采用二进制格式编码，采样频率为44.1kHz，量化为16位，这都是现代电子技术中光盘（CD）的标准格式。换句话说，这意味着，一旦给出了初始位模式，则通过将1-2-3 CA的规则顺序应用到初始位模式来完全再现整个数据。此外，描述给定数据所需的数据总量少于原始数据。因此，我们的方法可以在不丢失信息的情况下进行数据压缩，这是本项目中提出的一个新颖的想法。为了更详细地研究这个话题，我们于2000年11月举办了一次小型研讨会，举办地点位于冈山大学附近。受邀演讲嘉宾有（1）和田正人；广岛大学(2) 有马贵之；冈山大学(3) Song-Ju Kim ;早稻田大学(4) 永井嘉德;早稻田大学(5) 相泽洋司;早稻田大学其他参加者还有黑岩仗介（广岛大学）、奈良重利（冈山大学）以及冈山大学的几位研究生。

项目成果

S.Nakayama,J.Kuroiwa & S.Nara: "Partly Inverted Synaptic Connections & Complex Dynamics in a Symmetric Recurrent Neural Network Model"Proc.of Int.Conf.on Neural Information Processing (IC0NIP'2000). 2. 1274-1279 (2000)

中山，黑岩

S.Nakayama,J.Kuroiwa & S.Nara: "Partly Inverted Synaptic Connections & Complex Dynamics in a Symmetric Recurrent Neural Network Model"Proc.of Int.Conf.on Neural Information Processing (ICONIP'2000). 2. 1274-1279 (2000)

中山，黑岩

Yoshinori Nagai, et al.: "Discrete Algebra on Cellular Automata and Binary Textures"国士舘大学情報科学センター紀要. 22. 51-64 (2001)

Yoshinori Nagai 等人：“元胞自动机和二进制纹理的离散代数”，国士馆大学信息科学中心公告 22. 51-64 (2001)。

Yoshinori Nagai and Yoji Aizawa: "Rule-dynamical generalization of McCulloch-Pitts neuron networks"BioSystems. 58. 177-185 (2000)

Yoshinori Nagai 和 Yoji Aizawa：“McCulloch-Pitts 神经元网络的规则动力学概括”BioSystems。

Yoshinori Nagai and Yoji Aizawa: "Rule-dynamical generalization of McCulloch-Pitts neuron networks"BioSystems. 58. 177-185 (2000)

Yoshinori Nagai 和 Yoji Aizawa：“McCulloch-Pitts 神经元网络的规则动力学概括”BioSystems。