Both neurophysiological and psychophysical experiments have pointed out the crucial role of recurrent and feedback connections to process context-dependent information in the early visual cortex. While numerous models...
详细信息
Both neurophysiological and psychophysical experiments have pointed out the crucial role of recurrent and feedback connections to process context-dependent information in the early visual cortex. While numerous models have accounted for feedback effects at either neural or representational level, none of them were able to bind those two levels of analysis. Is it possible to describe feedback effects at both levels using the same model? We answer this question by combining Predictive coding (PC) and Sparse coding (SC) into a hierarchical and convolutional framework applied to realistic problems. In the Sparse Deep Predictive coding (SDPC) model, the SC component models the internal recurrent processing within each layer, and the PC component describes the interactions between layers using feedforward and feedback connections. Here, we train a 2-layered SDPC on two different databases of images, and we interpret it as a model of the early visual system (V1 & V2). We first demonstrate that once the training has converged, SDPC exhibits oriented and localized receptive fields in V1 and more complex features in V2. Second, we analyze the effects of feedback on the neural organization beyond the classical receptive field of V1 neurons using interaction maps. These maps are similar to association fields and reflect the Gestalt principle of good continuation. We demonstrate that feedback signals reorganize interaction maps and modulate neural activity to promote contour integration. Third, we demonstrate at the representational level that the SDPC feedback connections are able to overcome noise in input images. Therefore, the SDPC captures the association field principle at the neural level which results in a better reconstruction of blurred images at the representational level. Author summary One often compares biological vision to a camera-like system where an image would be processed according to a sequence of successive transformations. In particular, this "feedforward"
We have previously proposed Go-Back-i symbol automatic repeat request (GBi-ARQ) scheme appropriate to meteor burst communications. The GBi-ARQ scheme can achieve symbol-wise ARQ by using the Viterbi decoder for convol...
详细信息
ISBN:
(纸本)9784885522673
We have previously proposed Go-Back-i symbol automatic repeat request (GBi-ARQ) scheme appropriate to meteor burst communications. The GBi-ARQ scheme can achieve symbol-wise ARQ by using the Viterbi decoder for convolutional codes. In this paper, we propose a new carrier synchronization method appropriate to the GBi-ARQ scheme. The method can efficiently estimate carrier phase and frequency offset by using a pair of simple recursive equations for each survivor of the Viterbi algorithm.
Many-core architectures structured as fabrics of tightly-coupled clusters have shown promising results on embedded computer vision benchmarks, providing state-of-art performance with a reduced power budget. We propose...
详细信息
ISBN:
(纸本)9781479965885
Many-core architectures structured as fabrics of tightly-coupled clusters have shown promising results on embedded computer vision benchmarks, providing state-of-art performance with a reduced power budget. We propose PULP (Parallel processing Ultra-Low Power platform), an architecture built on clusters of tightly-coupled OpenRISC ISA cores, with advanced techniques for fast performance and energy scalability that exploit the capabilities of the STMicroelectronics UTB FD-SOI 28nm technology. As a use case for PULP, we show that a computationally demanding vision kernel based on convolutional Neural Networks can be quickly and efficiently switched from a low power, low frame-rate operating point to a high frame-rate one when a detection is performed. Our results show that PULP performance can be scaled over a 1x-354x range, with a peak performance/power efficiency of 211 GOPS/W.
Tactical communications is a communication system which is applied to a military operation with some strict requirements compared to a commercial communication system. One of the important requirement is its security....
详细信息
ISBN:
(纸本)9781479953035
Tactical communications is a communication system which is applied to a military operation with some strict requirements compared to a commercial communication system. One of the important requirement is its security. This communication has a high risk to be attacked by jamming which generated by the adversary. The other important requirement is network flexibility to handle its high mobility and information exchange in a big order. To fulfil those special requirement of tactical communication, OFDM system is equipped with channel coding and interleaving to handle information damage caused by jamming. The jamming strategy used in this research is single-tone and multitone jamming which strike while information signal pass the AWGN channel. In this research,we evaluate the performance of OFDM system which is equipped with convolutional Encoder and Inter leaver with various schemes. By using a half-rate convolutional encoder, for both parameters BER and interleaving gain, helical-scan interleaver scheme shows the best performance among block and random interleaver. However due to its complexity in applying this scheme, block interleaver which is the second best is preferred.
In embedded systems, many numerical algorithms are implemented with fixed-point arithmetic to meet area cost and power constraints. Fixed-point encoding decisions can significantly affect cost and performance. To eval...
详细信息
ISBN:
(纸本)9781479962785
In embedded systems, many numerical algorithms are implemented with fixed-point arithmetic to meet area cost and power constraints. Fixed-point encoding decisions can significantly affect cost and performance. To evaluate their impact on accuracy, designers resort to simulations. Their high running-time prevents thorough exploration of the design-space. To address this issue, analytical modeling techniques have been proposed, but their applicability is limited by scalability issues. In this paper, we extend these techniques to a larger class of programs. We use polyhedral methods to extract a more compact, graph-based representation of the program. We validate our approach with a several image and signal processing algorithms.
We propose here an algorithm for reconstructing an unknown constellation labeling. Our method assumes that the underlying error correcting code is a convolutional code. We define the notions of linear and affine equiv...
详细信息
ISBN:
(纸本)9784885522673
We propose here an algorithm for reconstructing an unknown constellation labeling. Our method assumes that the underlying error correcting code is a convolutional code. We define the notions of linear and affine equivalence among labelings. Those notions will help us to reduce the cost of the search. We show that the search is intractable with our method as the constellation size grows. In that case we restrict the search to Gray labelings. Our algorithm adapts very well to that constraint and allows an easy reconstruction up to a constellation of 256 points.
In what follows signal-code constructions employing block and convolutional nonbinary inner codes that can perform under severe jamming are proposed. Robustness of the proposed signal-code constructions to jamming is ...
详细信息
ISBN:
(纸本)9781479940509
In what follows signal-code constructions employing block and convolutional nonbinary inner codes that can perform under severe jamming are proposed. Robustness of the proposed signal-code constructions to jamming is verified by means of simulation.
A new method for spatial coupling of low-density parity-check ensembles is proposed. The method is inspired from overlapped layered coding. Edges of local ensembles and those defining the spatial coupling are separate...
详细信息
ISBN:
(纸本)9781479959990
A new method for spatial coupling of low-density parity-check ensembles is proposed. The method is inspired from overlapped layered coding. Edges of local ensembles and those defining the spatial coupling are separately built. The new method allows the construction of non-uniform coupling chains with near-Shannon spatially-varying thresholds under iterative decoding. A direct application of non-uniform spatial coupling is unequal error protection of information.
Low density parity check code (LDPC) has constructed a flexible, low decoding complexity and performance close to Shannon limit benefits. In this paper, the method of BICM with LDPC code as its component code is appli...
详细信息
ISBN:
(纸本)9781467363433
Low density parity check code (LDPC) has constructed a flexible, low decoding complexity and performance close to Shannon limit benefits. In this paper, the method of BICM with LDPC code as its component code is applied to the Differential Modulation OFDM (DM-OFDM) system, and the comparison with the traditional method of BICM with convolutional codes (CC-BICM) is analyzed. The simulation results show that, in the DM-OFDM system the performance of LDPC codes is better than the convolutional codes. And as a result of the use of non-coherent detection which does not require channel estimation, the complexity of system is greatly reduced.
In this paper, the performances of single carrier visible light communication systems are analyzed in multi-path fading optical channels. Distortions in the data sent due to the multipath optical channel are corrected...
详细信息
ISBN:
(纸本)9781728172064
In this paper, the performances of single carrier visible light communication systems are analyzed in multi-path fading optical channels. Distortions in the data sent due to the multipath optical channel are corrected by linear and nonlinear channel equalizers. Computer simulations are performed to compare the equalizers with each other and to obtain the uncoded and encoded bit error rate performances. As a result of the simulation results, it was observed that the convolutional coded and Turbo coded system eliminates the error floor at each iteration.
暂无评论