We present a generative factor analyzed hidden Markov model (GFA-HMM) for automatic speech recognition. In a traditional HMM, the observation vectors are represented by mixture of Gaussians (MoG) that are dependent on discrete-valued hidden state sequence. The GFA-HMM introduces a hierarchy of continuousvalued latent representation of observation vectors, where latent vectors in one level are a...

This paper presents the IfN’s Offline Handwritten Arabic Word Recognition System. The system uses Hidden Markov Models (HMM) for word recognition, and is based on character recognition without explicit segmentation. The first part of this paper deals with databases for word recognition systems, and in particular, the IFN/ENIT database. The second part gives a short description of the pre-proces...

Hidden Markov model (HMM) is a promising method that works well for images with variations in lighting, facial expression, and orientation. Face recognition draws attention as a complex task due to noticeable changes produced on appearance by illumination, facial expression, size, orientation and other external factors. To process images using HMM, the temporal or space sequences are to be cons...

5 The purpose of this note is to study the assumption of “mutual information inde6 pendence”, which is used by Zhou (2005) for deriving an output-dependent hidden 7 Markov model, the so-called discriminative HMM (D-HMM), in the context of deter8 mining a stochastic optimal sequence of hidden states. The assumption is extended 9 to derive its generative counterpart, the G-HMM. In addition, state...

5 The purpose of this note is to study the assumption of “mutual information inde6 pendence”, which is used by Zhou (2005) for deriving an output-dependent hidden 7 Markov model, the so-called discriminative HMM (D-HMM), in the context of deter8 mining a stochastic optimal sequence of hidden states. The assumption is extended 9 to derive its generative counterpart, the G-HMM. In addition, state...

Automatic musical accompaniment is where a human musician is accompanied by a computer musician. The computer musician is able to produce musical accompaniment that relates musically to the human performance. The accompaniment should follow the performance using observations of the notes they are playing. This paper describes a complete and detailed construction of a score following and accompa...

In this paper, an embedded isolated word recognition system, which is designed for command and control applications, is presented. The hardware is designed by using a 32 bit general purpose ARM7TDMI cored microcontroller and required peripherals. Vector Quantization (VQ) and Discrete Hidden Markov Model (HMM) based speech recognition algorithms are implemented on the embedded system. The system...

This paper employ the discrete hidden Markov model (HMM) in order to capture information about the Markov switching model’s inner states that is not directly observable, and to pre-detect the real estate business cycle’s volatility trend. The empirical results show that this HMM can capture the asymmetry in the duration of states. Compared with the real estate leading indicator announced by the...

The paper presents a complete discrete statistical framework, based on a novel vector quantization (VQ) front-end process. This new VQ approach performs an optimal distribution of VQ codebook components on HMM states. This technique that we named the distributed vector quantization (DVQ) of hidden Markov models, succeeds in unifying acoustic micro-structure and phonetic macro-structure, when th...