The clues for discontinuous structural transformation in handwriting lie in deformation around singular points and stroke connections of handwriting. For designing an algorithm for automatic construction of structural...
In this paper, we propose an integrated segmentation and recognition method for recognizing connected handwritten characters with recurrent neural network. It has been developed to both integrate segmentation and reco...
详细信息
A conceptual diagram is a kind of diagram which represents a logical structure among concepts using simple physical objects (loops, lines and character strings). This paper presents a method of interpretation of conce...
详细信息
In this paper a new effective feature-based resolution-based vector quantisation (VQ) method for Hermite-binomial transform domain image coding is proposed. Hermite-binomial transform is known to be highly relevant to...
详细信息
In this paper a new effective feature-based resolution-based vector quantisation (VQ) method for Hermite-binomial transform domain image coding is proposed. Hermite-binomial transform is known to be highly relevant to the human receptive field due to its association to the Gaussian derivative models. However, there are more transform coefficients than the original image samples due to the requirement of overlapped windows, which hindered its application to image coding. In the proposed scheme, we apply several VQ subcodebooks to encode image edge profiles and textures at different resolution levels. Simulation results on image coding showed that a high compression ratio can be obtained with good visual quality. Results have also been compared with that of JPEG image coding.
The paper presents an overview on architectures for VLSI implementations of video compression schemes as specified by standardization committees of the ITU and ISO, focussing on programmable architectures. Programmabl...
详细信息
The paper presents an overview on architectures for VLSI implementations of video compression schemes as specified by standardization committees of the ITU and ISO, focussing on programmable architectures. Programmable video signal processors are classified and specified as homogeneous and heterogeneous processor architectures. Heterogeneous processors outperform homogeneous processors due to adaptation to the requirements of special subtasks by dedicated modules. The majority of heterogeneous processors incorporate dedicated modules for high performance subtasks of high regularity like DCT and block matching. By normalization to a fictive 1.0 micron CMOS process typical linear relationships between silicon area and through-put rate have been determined for the different architectural styles. This relationship indicates a figure of merit for silicon efficiency.
Neural networks have been successfully applied to various pattern classification problems owing to their learning ability, high discrimination power, and excellent generalization ability. However, for the case of clas...
详细信息
This paper proposes a system that understands mathematical expressions on binarized printed document images. The system first extracts a set of »primitives» from the image. Each of the extracted primitives i...
详细信息
In this paper, the authors present a novel image coding technique which explicitly aims to retain perceptually important features of the image. The system exploits the psychovisual properties of the human visual syste...
详细信息
In this paper, the authors present a novel image coding technique which explicitly aims to retain perceptually important features of the image. The system exploits the psychovisual properties of the human visual system, in that it identifies perceptually important vector patterns (referred to as visual vectors) within small blocks of the image, and codes these using a scheme similar to vector quantization (VQ). Thus the system is referred to as visual vector quantization (VVQ). One of the main advantages of the new VVQ scheme over traditional VQ is that it is computationally much simpler, and yet its performance is comparable to that of traditional VQ. Experimental results are presented which demonstrate that the new technique can achieve visually satisfactory image reconstruction at a compression ratio as high as 17:1. Comparison of peak signal to noise ratio (PSNR) values shows that the performance of the new technique is comparable to that of other VQ-based image coders at a similar bit rate.
In this paper, we propose a new type of recurrent neural network for handwritten character recognition. The proposed recurrent neural network differs from Jordan and Elman recurrent neural networks in mew of functions...
详细信息
A method for low bit rate video coding based on wavelet lattice vector quantisation is proposed. By eliminating the blocking artefacts in the prediction error image introduced from the conventional block matching, the...
详细信息
A method for low bit rate video coding based on wavelet lattice vector quantisation is proposed. By eliminating the blocking artefacts in the prediction error image introduced from the conventional block matching, the overlapped block matching (OBM) motion compensation increases the efficiency of the video codec. The motion compensated prediction error signal is coded using Successive Approximation Wavelet Lattice Vector Quantisation (SAWLVQ), a method that combines the wavelet transform and lattice vector quantisation. Through comparison with the standard RMB model of the H-261 video codec, it is found that OBM-SAWLVQ codec results into improvements in both the peak signal-to-noise ratio performance and the subjective quality of the reconstructed pictures.
暂无评论