The Experts below are selected from a list of 237 Experts worldwide ranked by ideXlab platform
S.k. Mitra - One of the best experts on this subject based on the ideXlab platform.
-
Sensitivity analysis of low-complexity vector quantizers for focal-plane image compression
2004 IEEE International Symposium on Circuits and Systems (IEEE Cat. No.04CH37512), 2004Co-Authors: J.g.r.c. Gomes, S.k. MitraAbstract:Most high-performance block-coding systems for image compression, such as JPEG, have been designed for software or dedicated digital hardware implementations where the data are already assumed to be available in digital Format. In modern CMOS photosensors, smart-pixel technologies have allowed the realization of basic signal processing tasks at the pixel level, in Analog Format before Analog-to-digital (A/D) conversion. The elimination of A/D converters and implementation of block-coding directly over Analog blocks of pixels in such sensors can be attractive both in terms of area savings and power consumption. The design of block encoders, under the strong hardware constraints that derive from the A/D converter removal, has been investigated in this paper. We present a comparison of three systems in terms of rate, distortion and complexity, and present a numerical simulation analysis of their sensitivity to implementation errors. The conclusion of the analysis is that linear-transform coding vector quantizers outperform full-search vector quantizers and warping hyperbolic-tangent neural networks, in terms of performance, complexity and robustness, for a CMOS imaging sensor implementation.
-
ISCAS (5) - Sensitivity analysis of low-complexity vector quantizers for focal-plane image compression
2004 IEEE International Symposium on Circuits and Systems (IEEE Cat. No.04CH37512), 2004Co-Authors: J.g.r.c. Gomes, S.k. MitraAbstract:Most high-performance block-coding systems for image compression, such as JPEG, have been designed for software or dedicated digital hardware implementations where the data are already assumed to be available in digital Format. In modern CMOS photosensors, smart-pixel technologies have allowed the realization of basic signal processing tasks at the pixel level, in Analog Format before Analog-to-digital (A/D) conversion. The elimination of A/D converters and implementation of block-coding directly over Analog blocks of pixels in such sensors can be attractive both in terms of area savings and power consumption. The design of block encoders, under the strong hardware constraints that derive from the A/D converter removal, has been investigated in this paper. We present a comparison of three systems in terms of rate, distortion and complexity, and present a numerical simulation analysis of their sensitivity to implementation errors. The conclusion of the analysis is that linear-transform coding vector quantizers outperform full-search vector quantizers and warping hyperbolic-tangent neural networks, in terms of performance, complexity and robustness, for a CMOS imaging sensor implementation.
Hao Jiang - One of the best experts on this subject based on the ideXlab platform.
-
DAC - A spiking neuromorphic design with resistive crossbar
Proceedings of the 52nd Annual Design Automation Conference on - DAC '15, 2015Co-Authors: Chaofei Yang, Linghao Song, Yiran Chen, Qing Wu, Hai Li, Zheng Li, Hao JiangAbstract:Neuromorphic systems recently gained increasing attention for their high computation efficiency. Many designs have been proposed and realized with traditional CMOS technology or emerging devices. In this work, we proposed a spiking neuromorphic design built on resistive crossbar structures and implemented with IBM 130nm technology. Our design adopts a rate coding scheme where pre- and post-neuron signals are represented by digitalized pulses. The weighting function of pre-neuron signals is executed on the resistive crossbar in Analog Format. The computing result is transferred into digitalized output spikes via an integrate-and-fire circuit (IFC) as the post-neuron. We calibrated the computation accuracy of the entire system through circuit simulations. The results demonstrated a good match to our analytic modeling. Furthermore, we implemented both feedforward and Hopfield networks by utilizing the proposed neuromorphic design. The system performance and robustness were studied through massive Monte-Carlo simulations based on the application of digital image recognition. Comparing to the previous crossbar-based computing engine that represents data with voltage amplitude, our design can achieve >50% energy savings, while the average probability of failed recognition increase only 1.46% and 5.99% in the feedforward and Hopfield implementations, respectively.
-
A spiking neuromorphic design with resistive crossbar
2015Co-Authors: Qing Wu, Beiye Liu, Chenchen Liu, Bonan Yan, Chaofei Yang, Linghao Song, Yiran Chen, Hai Li, Zheng Li, Hao JiangAbstract:Neuromorphic systems recently gained increasing attention for their high computation efficiency. Many designs have been proposed and realized with traditional CMOS technology or emerging devices. In this work, we proposed a spiking neuromorphic design built on resistive crossbar structures and implemented with IBM 130nm technology. Our design adopts a rate coding scheme where pre- and post-neuron signals are represented by digitalized pulses. The weighting function of pre-neuron signals is executed on the resistive crossbar in Analog Format. The computing result is transferred into digitalized output spikes via an integrate-and-fire circuit (IFC) as the post-neuron. We calibrated the computation accuracy of the entire system through circuit simulations. The results demonstrated a good match to our analytic modeling. Furthermore, we implemented both feedforward and Hopfield networks by utilizing the proposed neuromorphic design. The system performance and robustness were studied through massive Monte-Carlo simulations based on the application of digital image recognition. Comparing to the previous crossbar-based computing engine that represents data with voltage amplitude, our design can achieve >50% energy savings, while the average probability of failed recognition increase only 1.46% and 5.99% in the feedforward and Hopfield implementations, respectively.
Surachet Kanprachar - One of the best experts on this subject based on the ideXlab platform.
-
WPMC - Spectral Vector Design for Gunfire Sound Classification System with a Smartphone using ANN
2018 21st International Symposium on Wireless Personal Multimedia Communications (WPMC), 2018Co-Authors: Settha Tangkawanit, Surachet KanpracharAbstract:In this research, a system for classifying the gunfire sound has been studied. The system is designed to function with a smartphone, which has a limited resource. The input sound is converted from the Analog Format to the digital one. The digital gunfire sound is then processed and analyzed in frequency domain using the smartphone. It is shown that, with noise injection method, using Artificial Neural Network (or ANN) in the classification process, the obtained accuracy for classifying 6 different gunfire sounds is considerably increased compared to the results found in [1]. Additionally, in this work, the feature vector with different number of bins in frequency domain is deeply studied. It is found that with a proper number of bins in the classification, the classification accuracy is significantly improved. The 100%-accuracy can be achieved for the SNR down to 10 dB and a very high accuracy (that is, greater than 90%) can be obtained at the 0-dB SNR. Using an appropriate number of feature vectors can lead to a promising performance in terms of gunfire sound classification.
-
Spectral Vector Design for Gunfire Sound Classification System with a Smartphone using ANN
2018 21st International Symposium on Wireless Personal Multimedia Communications (WPMC), 2018Co-Authors: Settha Tangkawanit, Surachet KanpracharAbstract:In this research, a system for classifying the gunfire sound has been studied. The system is designed to function with a smartphone, which has a limited resource. The input sound is converted from the Analog Format to the digital one. The digital gunfire sound is then processed and analyzed in frequency domain using the smartphone. It is shown that, with noise injection method, using Artificial Neural Network (or ANN) in the classification process, the obtained accuracy for classifying 6 different gunfire sounds is considerably increased compared to the results found in [1]. Additionally, in this work, the feature vector with different number of bins in frequency domain is deeply studied. It is found that with a proper number of bins in the classification, the classification accuracy is significantly improved. The 100%-accuracy can be achieved for the SNR down to 10 dB and a very high accuracy (that is, greater than 90%) can be obtained at the 0-dB SNR. Using an appropriate number of feature vectors can lead to a promising performance in terms of gunfire sound classification.
J.g.r.c. Gomes - One of the best experts on this subject based on the ideXlab platform.
-
Sensitivity analysis of low-complexity vector quantizers for focal-plane image compression
2004 IEEE International Symposium on Circuits and Systems (IEEE Cat. No.04CH37512), 2004Co-Authors: J.g.r.c. Gomes, S.k. MitraAbstract:Most high-performance block-coding systems for image compression, such as JPEG, have been designed for software or dedicated digital hardware implementations where the data are already assumed to be available in digital Format. In modern CMOS photosensors, smart-pixel technologies have allowed the realization of basic signal processing tasks at the pixel level, in Analog Format before Analog-to-digital (A/D) conversion. The elimination of A/D converters and implementation of block-coding directly over Analog blocks of pixels in such sensors can be attractive both in terms of area savings and power consumption. The design of block encoders, under the strong hardware constraints that derive from the A/D converter removal, has been investigated in this paper. We present a comparison of three systems in terms of rate, distortion and complexity, and present a numerical simulation analysis of their sensitivity to implementation errors. The conclusion of the analysis is that linear-transform coding vector quantizers outperform full-search vector quantizers and warping hyperbolic-tangent neural networks, in terms of performance, complexity and robustness, for a CMOS imaging sensor implementation.
-
ISCAS (5) - Sensitivity analysis of low-complexity vector quantizers for focal-plane image compression
2004 IEEE International Symposium on Circuits and Systems (IEEE Cat. No.04CH37512), 2004Co-Authors: J.g.r.c. Gomes, S.k. MitraAbstract:Most high-performance block-coding systems for image compression, such as JPEG, have been designed for software or dedicated digital hardware implementations where the data are already assumed to be available in digital Format. In modern CMOS photosensors, smart-pixel technologies have allowed the realization of basic signal processing tasks at the pixel level, in Analog Format before Analog-to-digital (A/D) conversion. The elimination of A/D converters and implementation of block-coding directly over Analog blocks of pixels in such sensors can be attractive both in terms of area savings and power consumption. The design of block encoders, under the strong hardware constraints that derive from the A/D converter removal, has been investigated in this paper. We present a comparison of three systems in terms of rate, distortion and complexity, and present a numerical simulation analysis of their sensitivity to implementation errors. The conclusion of the analysis is that linear-transform coding vector quantizers outperform full-search vector quantizers and warping hyperbolic-tangent neural networks, in terms of performance, complexity and robustness, for a CMOS imaging sensor implementation.
Settha Tangkawanit - One of the best experts on this subject based on the ideXlab platform.
-
WPMC - Spectral Vector Design for Gunfire Sound Classification System with a Smartphone using ANN
2018 21st International Symposium on Wireless Personal Multimedia Communications (WPMC), 2018Co-Authors: Settha Tangkawanit, Surachet KanpracharAbstract:In this research, a system for classifying the gunfire sound has been studied. The system is designed to function with a smartphone, which has a limited resource. The input sound is converted from the Analog Format to the digital one. The digital gunfire sound is then processed and analyzed in frequency domain using the smartphone. It is shown that, with noise injection method, using Artificial Neural Network (or ANN) in the classification process, the obtained accuracy for classifying 6 different gunfire sounds is considerably increased compared to the results found in [1]. Additionally, in this work, the feature vector with different number of bins in frequency domain is deeply studied. It is found that with a proper number of bins in the classification, the classification accuracy is significantly improved. The 100%-accuracy can be achieved for the SNR down to 10 dB and a very high accuracy (that is, greater than 90%) can be obtained at the 0-dB SNR. Using an appropriate number of feature vectors can lead to a promising performance in terms of gunfire sound classification.
-
Spectral Vector Design for Gunfire Sound Classification System with a Smartphone using ANN
2018 21st International Symposium on Wireless Personal Multimedia Communications (WPMC), 2018Co-Authors: Settha Tangkawanit, Surachet KanpracharAbstract:In this research, a system for classifying the gunfire sound has been studied. The system is designed to function with a smartphone, which has a limited resource. The input sound is converted from the Analog Format to the digital one. The digital gunfire sound is then processed and analyzed in frequency domain using the smartphone. It is shown that, with noise injection method, using Artificial Neural Network (or ANN) in the classification process, the obtained accuracy for classifying 6 different gunfire sounds is considerably increased compared to the results found in [1]. Additionally, in this work, the feature vector with different number of bins in frequency domain is deeply studied. It is found that with a proper number of bins in the classification, the classification accuracy is significantly improved. The 100%-accuracy can be achieved for the SNR down to 10 dB and a very high accuracy (that is, greater than 90%) can be obtained at the 0-dB SNR. Using an appropriate number of feature vectors can lead to a promising performance in terms of gunfire sound classification.
