Browsing by Author "Lin, RS"

Now showing 1 - 8 of 8

A modified morphological corner detector
(1998-03-01)
Lin, RS
;
Chu, CH
;
Hsueh, YC
;
資訊工程學系
;
Department of Computer Science
In this paper we propose a modified morphological corner detection method which finds convex and concave significant points using simple integer computation. We use the morphological peak extractor to detect convex corners and use a modified valley extractor to detect concave corners. Moreover, information of the distances between pixels on the boundary segment determined by two contiguous significant points and the chord connecting these two contiguous significant points can be applied to remedy the loosing corner due to the shape of the chosen structuring element. (C) 1998 Elsevier Science B.V. All rights reserved.
A new approach for audio classification and segmentation using Gabor wavelets and Fisher Linear Discriminator
(2005-09-01)
Lin, RS
;
Chen, LH
;
資訊工程學系
;
Department of Computer Science
Rapid increase in the amount of audio data demands an efficient method to automatically segment or classify audio stream based on its content. In this paper, based on the Gabor wavelet features, an audio classification and segmentation method is proposed. This method will first divide an audio stream into clips, each of which contains one-second audio information. Then, each clip is classified as one of two classes or five classes. Two classes contain speech and music; pure speech, pure music, song, speech with music background, and speech with environmental noise background are for five classes. Finally, a merge technique is provided to do segmentation. In order to make the proposed method robust for a variety of audio sources: we use Fisher Linear Discriminator to obtain features with the highest discriminative ability Experimental results show that the proposed method can achieve over 98% accuracy rate for speech and music discrimination, and more than 95% for a five-way discrimination. By checking the class types of adjacent clips, we can also identify more than 95% audio scene breaks in audio sequence.
A new approach for classification of generic audio data
(2005-02-01)
Lin, RS
;
Chen, LH
;
資訊工程學系
;
Department of Computer Science
The existing audio retrieval systems fall into one of two categories: single-domain systems that can accept data of only a single type (e.g. speech) or multiple-domain systems that offer content-based retrieval for multiple types of audio data. Since a single-domain system has limited applications, a multiple-domain system will be more useful. However, different types of audio data will have different properties, this will make a multiple-domain system harder to be developed. If we can classify audio information in advance, the above problems can be solved. In this paper, we will propose a real-time classification method to classify audio signals into several basic audio types such as pure speech, music, song, speech with music background, and speech with environmental noise background. In order to make the proposed method robust for a variety of audio sources, we use Bayesian decision function for multivariable Gaussian distribution instead of manually adjusting a threshold for each discriminator. The proposed approach can be applied to content-based audio/video retrieval. In the experiment, the efficiency and effectiveness of this method are shown by an accuracy rate of more than 96% for general audio data classification.
Content-based audio retrieval based on Gabor wavelet filtering
(2005-09-01)
Lin, RS
;
Chen, LH
;
資訊工程學系
;
Department of Computer Science
Rapid increase in the amount of audio data and especially music collections demand an efficient method to automatically retrieve audio objects based on its content. In this paper, based on the Gabor wavelet features, we will propose a method for content-based retrieval of perceptually similar music pieces in audio documents. It allows the user to select a reference passage within an audio file and retrieve perceptually similar passages such as repeating phrases within a music piece, similar music clips in a database or one song sung by different persons or in different languages. The proposed method will first divide an audio stream into clips, each of which contains one-second audio information. Then, the frame-based features of each clip are extracted based on the Gabor wavelet filters. Finally,, a similarity measuring technique is provided to perform pattern matching on the resulting sequences of feature vectors. Experimental results show that the proposed method can achieve over 96% accuracy rate for audio retrieval.
Efficient minimization of homogeneous FSMs for fault diagnosis
(1998-03-01)
Lin, RS
;
Yuang, MC
;
Hsu, SJ
;
資訊工程學系
;
Department of Computer Science
On the base of the Finite State Machine (FSM) model, the fault diagnosis problem deals with the identification of a faulty implementation against its specification represented by an FSM. In particular, to efficiently identify potential faulty implementations caused by a transfer error in an FSM, cross-verification over the FSM is performed after all potential faulty FSMs have been minimized. Existing minimization algorithms become inefficient due to the lack of taking advantage of the homogeneity of these potential faulty FSMs. In this paper, we propose a two-phase algorithm for the efficient and simultaneous minimization of a set of homogeneous faulty FSMs. Disregarding the suspicious transition, the first phase of the algorithm performs minimization of these faulty FSMs via a common digraph of th FSMs. These faulty FSMs are considered minimized if the distinguishing sequences of all state-pairs can be discovered from the common digraph. Otherwise, the algorithm in the second phase performs minimization for each faulty FSM by individually restoring its corresponding unexpected transition in the reduced common digraph. To demonstrate the efficiency of the two-phase algorithm, we carried out experiments on a number of realistic protocol specifications, including the Alternation Bit Protocol (ABP), Transport Protocol Class 4 (TP4), and ISDN Basic Rate Interface (BRI) protocol. Experimental results show that the algorithm renders the minimization complexity greatly reduced for most of the realistic protocol FSMs.
Multichannel filtering by gradient information
(2000-02-01)
Lin, RS
;
Hsueh, YC
;
資訊工程學系
;
Department of Computer Science
In this paper, multichannel image filtering using local gradient information is studied and evaluated which is simpler and more appropriate than the traditional approaches that have been addressed by means of groupwise vector ordering information. Two adaptive weighted multichannel filters based on local gradient information to noise removal in color images are introduced. The first proposed multichannel filter is rather highly effective for Gaussian and Uniform noise removal in preserving good edges. The second proposed multichannel filter is robust to Gaussian, Uniform, Impulse noise, and Gaussian noise mixed with outliers. (C) 2000 Elsevier Science B.V. All rights reserved.
Some modifications of gradient weighted filters
(1999-12-01)
Lin, RS
;
Hsueh, YC
;
資訊工程學系
;
Department of Computer Science
Gradient weighted filters are locally adaptive weighted mean filters. In this paper, a general formulation of gradient weighted filters with some characteristic parameters was derived first from existing gradient weighted filters. Then we propose some modifications by varying these parameters. We modify gradient inverse weighted filters, characterize filters into first and second order filters, and propose Pi filters. Moreover, an imposed criterion for second order filters to preserve fine details was introduced to promote the existing gradient weighted filters. Finally, a criterion to combine first and second order filters was proposed to remove noise with mixed types, Throughout this paper, rational analysis and experimental results demonstrate the efficiency of the proposed methods. (C) 1999 Academic Press.
Synchronization paradigm for protocol testing under multiparty configuration
(1999-04-01)
Lin, RS
;
Yuang, MC
;
資訊工程學系
;
Department of Computer Science
Protocol testing leads to the synchronization problem should test sequences be applied to multiple distanced testers, namely under the multiparty configuration. This paper presents a novel synchronization paradigm which seamlessly unifies two synchronization techniques, self-synchroniable sequences and external synchronization operations, by means of the state expansion transformation. In the paradigm, the protocol specification is first transformed into a state expansion digraph with two pieces of datum augmented. They are: (a) a zero cost assigned to each synchronization-problem-free crossing from one transition to another transition, and (b) a weighted cost assigned to each external synchronization operation whenever the synchronization is deemed necessary. On the basis of the state expansion transformation, synchronizable, optimal sequences for testing can be efficiently derived. To demonstrate the viability of the proposed paradigm, we present the generations of two synchronizable sequences, namely the synchronizable preamble and the synchronizable distinguishing sequence, which have previously been used for the testing of the correctness of a protocol's transition. The paper also shows that the complexities of the two sequences generations are also polynomial-bounded, i.e., O((np(2))log(np)) and O((n(2)p(2)) log(np)), respectively, where n and p are the numbers of states and input symbols of the protocol specification. (C) 1999 Elsevier Science Ltd. All rights reserved.