Video encoders and decoders for HEVC-like compression standards require huge external memory bandwidth, which occupies a significant portion of the codec power consumption. To reduce this paper presents new lossless reference frame recompression algorithm along with high-throughput hardware architecture. Firstly, hybrid spatial -domain prediction is proposed to combine merits DPCM scanning averaging. The then enhanced multiple modes accommodate various image characteristics. Finally,...

Approximate computing is applicable to improve hardware performance by sacrificing some accuracy for error-tolerant applications, where multiplication a key arithmetic operation. In this paper, we propose low-cost approximate multiplier design employing new probability-driven inexact 4:2, 6:2, 8:2 compressors and half-adders. This compressor explored reduce the height of partial product matrix into two rows. Different levels can be achieved through grouped error recovery scheme that employs...

Computer vision applications are rapidly gaining popularity in embedded systems, which typically involve a difficult tradeoff between performance and energy consumption under constraint of real-time processing throughput. Recently, hardware (FPGA ASIC-based) implementations have emerged, significantly improves the efficiency computation. These implementations, however, often intensive memory traffic that retains significant portion at system level. To address this issue, we first researchers...

Motion estimation and motion compensation in HEVC similar video codecs involve huge memory traffic storing loading reference frames. The resulting power composes a significant portion of system energy consumption. This paper presents reduction framework that losslessly compresses decompresses frames on-the-fly. We first present the architecture supports random access frame data compressed variable ratios. latest recompression algorithms corresponding VLSI implementation are also introduced....

This paper presents an angular measurement intra prediction algorithm compatible with compressive sensing (CS) based image sensors. We propose to apply the idea of H.264 and emulate its computation in compression. More structural rows random 0/1 matrix are designed for embedding more precise boundary information neighboring into measurements prediction. The proposed that can be directly applied captured by sensor. Experiment results show could increase coding efficiency, 10% BD-rate...

This paper presents a measurement-domain intra prediction coding framework that is compatible with compressive sensing (CS) based image sensors. In this framework, we propose low-complexity algorithm can be directly applied to the measurements captured by sensor. Moreover, structural random 0/1 measurement matrix, embedding block boundary information extracted from for prediction. Experiment results show our proposed compress and increase efficiency, 30% BD-rate reduction compared direct...

This paper presents a measurement-domain intra prediction coding framework that is compatible with compressive sensing (CS)-based image sensors. In this framework, we propose low-complexity algorithm can be directly applied to measurements captured by the sensor. We proposed structural random 0/1 measurement matrix, embedding block boundary information extracted from for prediction. Furthermore, low-cost Very Large Scale Integration (VLSI) architecture implemented substituting matrix...

FRC (frame re-compression) is a kind of widely used technique in reducing the SDRAM (synchronous dynamic random access memory) bandwidth HEVC video system. However, previous research works, imposes requirements on accessing pattern and hence its usage are only limited codecs. While typical VLSI system, there exists many other IPs with high requirements. Therefore, this article, we propose new architecture to overcome limitation make it applicable all which raises overall reduction rate whole...

For mobile video codecs, the huge energy dissipation for external memory traffic is a critical challenge under battery power constraint. Lossy embedded compression (EC), as solution to this challenge, considered in paper. While previous studies EC mostly focused on algorithms at block level, work, best of our knowledge, first one that addresses allocation quality and frame level. lossy EC, main difficulty its application lies error propagation from degradation reference frames....

Computer vision applications are rapidly gaining popularity in embedded systems, which typically involve a difficult trade-off between performance and energy consumption under constraint of real-time processing throughput. Recently, hardware (FPGA ASIC-based) implementations have emerged that significantly improve the efficiency computation. These implementations, however, often intensive memory traffic retains significant portion at system level. To address this issue, we present lossy...

For mobile video codecs, the huge energy dissipation for external memory traffic is a critical challenge under battery power constraint. Lossy embedded compression (EC), as solution to this challenge, considered in paper. While previous studies EC mostly focused on algorithms at block level, work, best of our knowledge, first one that addresses allocation quality and frame level. lossy EC, main difficulty its application lies error propagation from degradation reference frames....

For mobile video codecs, the huge energy dissipation for external memory traffic is a critical challenge under battery power constraint. Lossy embedded compression (EC), as solution to this challenge, considered in paper. While previous studies lossy EC mostly focused on algorithm optimization reduce distortion, work, best of our knowledge, first one that addresses distortion control. Firstly, from both theoretical analysis and experiments optimization, conclusion drawn that, at frame level,...

Convolutional neural networks (CNNs) are rapidly gaining popularity in artificial intelligence applications and employed mobile devices. However, this is challenging because of the high computational complexity CNNs limited hardware resource To address issue, compressing CNN model an efficient solution. This work presents a new framework compression, with sparseness ratio allocation (SRA) neuron re-pruning (NRP). achieve higher overall spareness ratio, SRA exploited to determine pruned...