Browsing by Author "Xu, Yeong-Yuh"

Now showing 1 - 3 of 3

A multiple-instance neural networks based image content retrieval system
(2006)
Chuang, Shun-Chin
;
Xu, Yeong-Yuh
;
Fu, Hsin Chia
;
Huang, Hsiang-Cheh
;
資訊工程學系
In this paper, we proposed a Multiple-Instance Neural Network (MINN) for content-based image retrieval (CBIR). In order to represent the rich content of an image without precisely image segmentation, the image retrieval problem is considered as a multiple-instance learning problem. A set of exemplar images are selected by a user, each of which is labelled as conceptual related (positive) or conceptual unrelated (negative) image. Then, the proposed MINN is trained by using the proposed learning algorithm to learn the user's preferred image concept from the positive and negative examples. Experimental results show that: (1) without image segmentation and using only the color histogram as the image feature, the MINN without relevance feedback performs slightly inferior to some leading image retrieval methods, and (2) the MINN with the relevance feedback can significantly improve the retrieving performance from 40.3% to 59.3%, which outperforms to the results of some leading image retrieval methods.
Automatic construction of regression class tree for MLLR via model-based hierarchical clustering
(2006)
Cheng, Shih-Sian
;
Xu, Yeong-Yuh
;
Wang, Hsin-Min
;
Fu, Hsin-Chia
;
資訊工程學系
In this paper, we propose a model-based hierarchical clustering algorithm that automatically builds a regression class tree for the well-known speaker adaptation technique - Maximum Likelihood Linear Regression (MLLR). When building a regression class tree, the mean vectors of the Gaussian components of the model set of a speaker independent CDHMM-based speech recognition system are collected as the input data for clustering. The proposed algorithm comprises two stages. First, the input data (i.e., all the Gaussian mean vectors of the CDHMMs) is iteratively partitioned by a divisive hierarchical clustering strategy, and the Bayesian Information Criterion (BIC) is applied to determine the number of clusters (i.e., the base classes of the regression class tree). Then, the regression class tree is built by iteratively merging these base clusters using an agglomerative hierarchical clustering strategy, which also uses BIC as the merging criterion. We evaluated the proposed regression class tree construction algorithm on a Mandarin Chinese continuous speech recognition task. Compared to the regression class tree implementation in HTK, the proposed algorithm is more effective in building the regression class tree and can determine the number of regression classes automatically.
Texture recognition by generalized probabilistic decision-based neural networks
(2011-05-01)
Xu, Yeong-Yuh
;
Tseng, C. -L.
;
Fu, Hsin-Chia
;
資訊工程學系
;
Department of Computer Science
Texture recognition have received tremendous attentions in the past decades, due to its wide applications in computer vision and pattern recognition. For various applications, formulating texture features in distributional forms can sometimes provide meaningful representation than in numerical forms. In this paper, a generalized probabilistic decision-based neural network (GPDNN), based on a novel methodology for the measurement of the difference between two distributions, is proposed for texture recognition. Based on a two-layer pyramid-type network structure, the proposed GPDNN receives texture data via 2-D grid input nodes, and outputs the classification and/or retrieval results at the top layer node. Our prototype system demonstrates a successful utilization of GPDNN to the texture recognition on 40 texture images selected from the MIT Vision Texture (VisTex) database. Regarding the performance, experiment results show that (1) based on the proposed distribution difference measurement method, the texture retrieval accuracy is improved from 77% to 82% by comparing with some recently published leading methods, and (2) the proposed GPDNN has significant improvements in classification accuracy from 82.2% to 90.1% and retrieval accuracy from 79.9% to 88.6% by comparing with traditional approaches. (C) 2010 Elsevier Ltd. All rights reserved.