Yoshihiro Sugaya

  • Citations Per Year
Learn More
This paper addresses the automatic generation of a typographic font from a subset of characters. Specifically, we use a subset of a typographic font to extrapolate additional characters. Consequently, we obtain a complete font containing a number of characters sufficient for daily use. The automated generation of Japanese fonts is in high demand because a(More)
A fuzzy inference model for learning from experiences (FILE) is proposed. the model can learn from experience data obtained by trial-and-error of a task and it can stably learn from both experiences of success and failure of a trial. the learning of the model is executed after each of trial of the task. hence, it is expected that the achievement rate(More)
Fuzzy inference models can conduct advanced inference using knowledge which is easily understood by humans. In this paper, we propose a leaning fuzzy inference model. The model can learn with experience data obtained by trial-and-error of a task. The learning of the model is executed after each trial of the task. Hence, it is expected that the achievement(More)
本論文では,記憶容量制限下で時系列を逐次学習可能な,適応的逐次学習ネットワークモデル (Adaptive and Sequential Learning Network model : ASLN model) を提案する.提案モデルは入力として与えられる時系列情 報の逐次学習が可能であり,記憶の忘却を行うことで記憶容量の有効利用を図りながら,周囲の環境の変化に適応的 に知識を獲得する.モデルは入力時系列に含まれる要素と,その遷移に関する情報を記憶する.モデルは時系列の要 素のなかでも入力頻度の高いものを重要と判断し,優先的に記憶する.また入力される要素の遷移に関する情報を状 態ベクトルという形で表現し,階層型ニューラルネットワークにより記憶する.状態ベクトルは過去の入力系列の情(More)
This paper describes an adaptive deblocking postfilter based on neural networks for use in H.265 High Efficiency Video Coding (HEVC). Blocking noise is a common problem in video coding caused by the division of the frame into blocks. The filter is adaptive because it uses different filter parameters depending on block characteristics. We use a modified HEVC(More)
In character recognition system using images captured by a mobile phone camera, it is difficult to correctly recognize blurred characters. Therefore, we aim at high accuracy character recognition by obtaining information from multiple frames of video. In this paper, we propose a character recognition method based on the Compound Mutual Subspace Method which(More)
Intelligent systems using conventional associative memory model learn the knowledge which is prepared in advance. Therefore they have poor adaptability to the environment different from what the knowledge assumes.In this study, we propose an associative memory model to acquire procedural knowledge sequentially from surroundings of the model.We carried out(More)