Chih-Hsun Chou

Automatic recognition of animal vocalizations using averaged MFCC and linear discriminant analysis

In this paper we propose a method that uses the averaged Mel-frequency cepstral coefficients (MFCCs) and linear discriminant analysis (LDA) to automatically identify animals from their sounds. First, each syllable corresponding to a piece of vocalization is segmented. The averaged MFCCs over all frames in a syllable are calculated as the vocalization features. Linear discriminant analysis (LDA), which finds out a transformation matrix that minimizes the within-class distance and maximizes the between-class distance, is utilized to increase the classification accuracy while to reduce the dimensionality of the feature vectors. In our experiment, the average classification accuracy is 96.8% and 98.1% for 30 kinds of frog calls and 19 kinds of cricket calls, respectively.

Continuous Birdsong Recognition Using Gaussian Mixture Modeling of Image Shape Features

Traditional birdsong recognition approaches used acoustic features based on the acoustic model of speech production or the perceptual model of the human auditory system to identify the associated bird species. In this paper, a new feature descriptor that uses image shape features is proposed to identify bird species based on the recognition of fixed-duration birdsong segments where their corresponding spectrograms are viewed as gray-level images. The MPEG-7 angular radial transform (ART) descriptor, which can compactly and efficiently describe the gray-level variations within an image region in both angular and radial directions, will be employed to extract the shape features from the spectrogram image. To effectively capture both frequency and temporal variations within a birdsong segment using ART, a sector expansion algorithm is proposed to transform its spectrogram image into a corresponding sector image such that the frequency and temporal axes of the spectrogram image will align with the radial and angular directions of the ART basis functions, respectively. For the classification of 28 bird species using Gaussian mixture models (GMM), the best classification accuracy is 86.30% and 94.62% for 3-second and 5-second birdsong segments using the proposed ART descriptor, which is better than traditional descriptors such as LPCC, MFCC, and TDMFCC.

Bird Species Recognition by Comparing the HMMs of the Syllables

In this study, a bird species recognition system based on their sounds is proposed. In this system, the birdsong of a bird species is segmented into many syllables, from which several primary frequency sequences can be obtained. By using the statistics of the principle frequency sequences, all the syllables are clustered with the fuzzy C-mean clustering method so that each syllable group can be modeled by a hidden Markov model (HMM) characterizing the features of the song of the bird species. Using the Viterbi algorithm, the recognition process is achieved by finding the template bird species that has the most probable HMMs matching the frequency sequences of the test birdsong. Experimental results show that the proposed system can achieve a recognition rate of over 78% for 420 kinds of bird species.

GA based optimal keyword extraction in an automatic chinese web document classification system

The main steps for designing an automatic document classification system include feature extraction and classification. In this paper a method to improve feature extraction is proposed. In this method, genetic algorithm (GA) was applied to determine the threshold values of four criteria for extracting the representative keywords for each class. The purpose of these four threshold values is to extract as few representative keywords as possible. This keyword extraction method was combined with two classification algorithms, vector space model (VSM) and support vector machine (SVM), for examining the performance of the proposed classification system under various extracting conditions.

GA-Based Keyword Selection for the Design of an Intelligent Web Document Search System

The main steps for designing an automatic document classification system include feature extraction and classification. In this article a method to improve feature extraction is proposed. In this method, genetic algorithm was applied to determine the threshold values of four criteria for extracting the representative keywords for each class. The purpose of these four threshold values is to extract as few representative keywords as possible. This keyword extraction method was combined with two classification algorithms, vector space model and support vector machine, for examining the performance of the proposed classification system under various extracting conditions.

Modulation Spectral Analysis of Static and Transitional Information of Cepstral and Spectral Features for Music Genre Classification

In this paper, we will propose an automatic music genre classification approach based on long-term modulation spectral analysis on the static and transitional information of spectral (OSC and MPEG-7 NASE) as well as cepstral (MFCC) features. An information fusion approach which integrates both feature level fusion and decision level combination is employed to improve the classification accuracy. Experiments conducted on the music database employed in the ISMIR2004 Audio Description Contest have shown that the proposed approach can achieve a classification accuracy of 87.79%, which is better than the winner of the contest.

Music genre classification using modulation spectral features and multiple prototype vectors representation

In this paper, we will propose an automatic music genre classification approach based on long-term modulation spectral analysis of spectral (OSC and MPEG-7 NASE) as well as cepstral (MFCC) features. A modulation spectrogram corresponding to the collection of modulation spectra of MFCC/OSC/NASE will be constructed. The modulation spectrum is then decomposed into several logarithmically spaced modulation subbands. For each modulation subband, a new set of modulation spectral features, including modulation spectral contrast (MSC), modulation spectral valley (MSV), modulation spectral energy (MSE), modulation spectral centroid (MSCEN) and modulation spectral flatness (MSF) are then computed from each modulation subband. To cope with the problem that the feature vectors extracted from the music tracks of identical music genre might differ significantly, each music genre is modeled with a number of representative prototype vectors generated by c-means clustering algorithm. An information fusion approach which integrates both feature level fusion method and decision level combination method is then employed to improve the classification accuracy. Experiments conducted on ISMIR 2004 music dataset have shown that our proposed approach can achieve higher classification accuracy than other approaches with the same experimental setup.

3D model retrieval using 2D cepstral features

In this paper, we will propose a 3D model retrieval approach using 2D cepstral features. First, six projection planes representing the elevation (depth) value are generated. Then, 2D cepstral features are extracted from each projection plane for searching similar 3D models. Experiments conducted on the Princeton Shape Benchmark (PSB) database have shown that the proposed 2D cepstral features outperforms other state-of-the-art descriptors in terms of the DCG score.

A 3D Model Retrieval System Based on the Cylindrical Projection Descriptor

In recent years, the demand for a content-based 3D model retrieval system becomes an important issue. In this paper, the cylindrical projection descriptor (CPD) will be proposed for 3D model retrieval. To derive better retrieval results, the CPD will be combined with the radial distance descriptor (RDD). The experiments are conducted on the Princeton Shape Benchmark (PSB) database. Experiment results show that our proposed method is superior to others.

Fast Forgery Detection with the Intrinsic Resampling Properties

With the rapid progress of the image editing software, the image forgery can leave no visual clues on the tampered regions and makes us unable to judge the image authenticity. In general, the digital image forgery often utilizes the scaling, rotation or skewing operations in which the resampling and interpolation processes are demanded. By observing the detectable periodic properties introduced from the resampling and interpolation processes, we propose a novel method based on the intrinsic properties of resampling scheme to detect the tampered regions with the pre-calculated resampling weighting table and the periodic properties of prediction error distribution. The experimental results show that the proposed method outperforms the conventional methods in terms of efficiency and accuracy.