Hisashi Aomori

Sigma-delta cellular neural network for 2D modulation

Although sigma-delta modulation is widely used for analog-to-digital (A/D) converters, sigma-delta concepts are only for 1D signals. Signal processing in the digital domain is extremely useful for 2D signals such as used in image processing, medical imaging, ultrasound imaging, and so on. The intricate task that provides true 2D sigma-delta modulation is feasible in the spatial domain sigma-delta modulation using the discrete-time cellular neural network (DT-CNN) with a C-template. In the proposed architecture, the A-template is used for a digital-to-analog converter (DAC), the C-template works as an integrator, and the nonlinear output function is used for the bilevel output. In addition, due to the cellular neural network (CNN) characteristics, each pixel of an image corresponds to a cell of a CNN, and each cell is connected spatially by the A-template. Therefore, the proposed system can be thought of as a very large-scale and super-parallel sigma-delta modulator. Moreover, the spatio-temporal dynamics is designed to obtain an optimal reconstruction signal. The experimental results show the excellent reconstruction performance and capabilities of the CNN as a sigma-delta modulator.

Separable 2D Lifting Using Discrete-Time Cellular Neural Networks for Lossless Image Coding

The lifting scheme is an efficient and flexible method for the construction of linear and nonlinear wavelet transforms. In this paper, a novel lossless image coding technique based on the lifting scheme using discrete-time cellular neural networks (DT-CNNs) is proposed. In our proposed method, the image is interpolated by using the nonlinear interpolative dynamics of DT-CNN, and since the output function of DT-CNN works as a multi-level quantization function, our method composes the integer lifting scheme for lossless image coding. Moreover, the nonlinear interpolative dynamics by A-template is used effectively compared with conventional CNN image coding methods using only B-template. The experimental results show a better coding performance compared with the conventional lifting methods using linear filters.

Hybrid lifting scheme using discrete-time cellular neural networks for lossless image coding

The lifting scheme is a flexible method for the construction of linear and nonlinear wavelet transforms. In the nonlinear lifting scheme, it is difficult to design the optimal update filter corresponding to the nonlinear prediction filter. The hybrid use of the linear filter and the nonlinear filter is an efficient method for obtaining an optimal filter pair. In this paper, we propose a novel hybrid lifting scheme using discrete-time cellular neural networks (DT-CNNs) for lossless image coding. In our method, the image is interpolated by using the nonlinear interpolative dynamics of DT-CNNs, and the update process of lifting is designed by using the linear 5-tap filter to avoid the aliasing. Since the output function of DT-CNNs works as a multilevel quantizing function, our method composes the integer lifting scheme for lossless image coding. Moreover, our method makes good use of the nonlinear interpolative dynamics by A-template compared with conventional CNN image coding methods using only B-template. The experimental results show a better coding performance compared with those of the conventional lifting method using linear filters.

Lossless image coding by cellular neural networks with minimum coding rate learning

In this paper, a novel lossless image coding scheme using the cellular neural network (CNN) is proposed. From the viewpoint of the optimal lossless coding, our method is optimized for not only mean squared error (MSE) but also a coding rate. The key idea of this work is that the local structure of an image is modeled by six types of CNN templates in order to achieve high prediction performance, and the CNN parameters that gives prediction characteristic are decided by the supervised minimum coding rate learning. Moreover, in the entropy coding layer, the prediction residuals are coded by an adaptive arithmetic encoder with context modeling for high coding efficiency. The effectiveness of proposed algorithm is confirmed by some computer simulations of various standard test images, and its performance is compared with that of conventional hierarchical coding schemes having scalability.

A Spatial Domiain Sigma-Delta Modulator Using Discrete-Time Cellular Neural Networks

In this paper, a novel spatial domain sigma-delta modulator using discrete-time cellular neural networks (DT-CNNs) is proposed. Since the nature of CNN dynamics with the output function which has two saturation regions is to binarize the input image, CNNs have a capabilities as a spatial domain sigma-delta modulator. In the proposed architecture, the A-template is used for a digital to analogue converter (DAC), the C-template works as an integrator, and the nonlinear output function is for the bilevel output. Moreover, the dynamics is designed for obtaining an optimal reconstruction image. The experimental results show a good reconstruction performance and capabilities of CNN as a sigma-delta modulator

Missing image interpolation using sigma-delta modulation type of DT-CNN

This paper proposes a new interpolation method for an incomplete image using sigma-delta modulation type of Discrete-Time Cellular Neural Networks. Missing pixels in an image are interpolated by function of its nearest values using B-template with Gaussian filter. We can reconstruct analog image which has missing values into digital image by using this framework. We evaluated our new proposed method with six standard images which have missing pixels at various percentages of missing values. The experimental results show that, by using sigma-delta modulation type of Discrete-Time Cellular Neural Networks, we can achieve a high peak signal-to-noise ratio for various image datasets and at different rates of missingness.

Design of DT-CNN for Imputing Data at Unobserved Location of Geostatistics Image Dataset

The presence of missing values in a geostatistics dataset can affect the performance of using those dataset as generic purposed. In this paper, we have developed a novel method to estimate missing observation in geostatistics by using sigma-delta modulation type of Discrete-Time Cellular Neural Networks(DT-CNN). The nearest neighboring pixels of missing values in an image are used. The interpolation process is done by using B-template with Gaussian filter. The DT-CNN is used for reconstructing the imputed values from analog image value to digital image value. We have evaluated this approach through the experiments on geostatistics image which has different characteristics of missing pixels such as Landsat 7 ETM+ SLC-off and standard geostatistics image. The experimental results show that by using sigma-delta modulation type of Discrete-Time Cellular Neural Networks, we can achieve a high PSNR for various image datasets and at different characteristics of missing image.

An oversampling 2D sigma-delta converter by cellular neural networks

The sigma-delta cellular neural network (SD-CNN) is a novel framework of spatial domain sigma-delta modulation utilizing neuro dynamics. Also, it has signal reconstruction and noise shaping characteristics that are important sigma-delta properties. Although the noise shaping effect with the oversampling technique plays very important role for drastic quantization noise reduction in binary digital sequences, the conventional SD-CNN could not use it effectively since it can be thought that the time-domain and spatial-domain oversampling are effective for the SD-CNN. In this paper, a novel SD-CNN with the oversampling technique for an analogue DC input is proposed. Experimental results of various standard test images in several oversampling ratios suggest that the proposed oversampling SD-CNN has an excellent AD and DA performance.

Hierarchical lossless image coding using cellular neural network

In this paper, a novel hierarchical lossless image coding scheme using the cellular neural network (CNN) is proposed. The coding architecture of the proposed method is based on the lifting scheme that is one of the scalable coding framework for still images, and its coding performance strongly depends on the prediction ability. To cope with this spontaneously characteristic, an image interpolation is modeled by an optimal problem that minimizes the prediction error. To achieve the high accuracy prediction with a low coding rate, two types of templates are used for dealing with the local structure of the image, and the CNN parameters are decided by the minimum coding rate learning. In the coding layer, the arithmetic coder with context modeling is used for obtaining a high coding efficiency. Experimental results in various standard test images suggest that the coding performance of our proposed method is better than that of conventional hierarchical coding schemes.

A Spatial Domain Sigma-Delta Modulation via Discrete-Time Cellular Neural Networks

In this paper, a novel spatial domain sigma-delta modulation using two-layered discrete-time cellular neural networks (DT-CNNs) is proposed. Since the nature of CNN dynamics with the output function which has two saturation regions is to binarize the input image, the dynamics has a capabilities for a digital image halftoning. In the proposed architecture, the nonlinear interpolative dynamics is exploited to obtain an optimal reconstruction image from the bilevel modulated image, and quantization noises are spatially distributed by the noise shaping property of the dynamics. The experimental results show a excellent reconstruction performance and capabilities of the CNN as a sigma-delta modulation.