A low power seizure detection processor based on direct use of compressively-sensed data and employing a deterministic random matrix

Jafari, Ali; Page, Adam; Sagedy, Chris; Smith, Emily A.

doi:10.1109/biocas.2015.7348376

Cited by 21 publications

(9 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Undoubtedly, in this particular strategy, there is actually no demand to spare all nos as well as ones in the ROM. In previous job [1], the writers that the mapping out construction utilizing DRM strategy reduces the equipment cost matched up to outlining element making use of LFSR circuit [8], [9]. Fig.…”

Section: A Sketching Algorithmmentioning

confidence: 99%

“…The efficiency of the confiscation sensor is actually defined in regards to sensitiveness and also uniqueness. Level of sensitivity pertains to the percent of confiscation beginnings recognized and also Uniqueness pertains to the portion of improperly sensed confiscation starts [1]. Likewise, to predict the sensor efficiency on the data coming from an individual, a leave-one record-out cross-validation method is actually made use of [12] The suggested device could possibly attain a sensitiveness of 81.8% as well as the uniqueness of 93.9% for Nyquist-domain confiscation data (without Outlining).…”

Section: Bmentioning

confidence: 99%

“…To manipulate concealed data patterns and also accomplish effective forecasts, Artificial intelligence (ML) strategies are taken on. As an example, in clever as well as wearable wellness surveillance tools, ML methods are embraced to forecast future carcinogen such as boosted high blood pressure, blood-glucose degrees, drop discovery as well as seizure discovery [1], [2], [3] To offer reliable wellness keeping track of a multi-sensor strategy is embraced, where each sensor generates multitude of examples per secondly. At every example duration, keeping track of tool requires to refine a number of DSP and also ML bits.…”

Section: Introductionmentioning

confidence: 99%

“…Feature extraction and also distinction is actually carried out regionally at the sensor [1] Hence, it transfers a single little bit prophecy i.e confiscation or even no-seizure, this lowers gearbox energy dramatically. Furthermore, the transmitter may be in the sleeping method while processing.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Programmable Big Data Processing Framework to Reduce On-Chip Communications and Computations Towards Reducing Energy of the Processing

Medavaka¹

2019

International Journal of Advanced Research in Computer and Comm

View full text Add to dashboard Cite

Multi-Sensor health and wellness tracking systems are utilized to forecast future occasions of our health and wellness system. Each sensor produces a whopping quantity of data per 2nd and also requires to be refined in real-time. At the very same time, health and wellness tracking systems are battery ran, hence they have stiff restraints on power and also the location of the processing platform. Furthermore, health and wellness surveillance systems need to be precise, hence we adjust artificial intelligence methods to boost discovery precision. We recommend a programmable Big Data Processing structure to decrease on-chip interactions as well as calculations, therefore minimizing the power of the processing. We incorporate a low-overhead laying out structure with a low-power programmable PENC many-core platform. The laying out method decreases the data interactions and also calculations, furthermore processing time is reduced by parallel processing on the many-core platform. For the presentation we reveal seizure discovery application with 22-channel of electroencephalograph (EEG), each network creates 256 examples per 2nd calling for overall of 88 Kbps data price. The calculations are lowered by 16 while the power intake of processing is minimized approximately 68%. For compression prices of 2-16, the seizure discovery efficiency for the level of sensitivity and also uniqueness is broken down by 2.07% and also 2.97%, specifically for Logistic Regression classifier.

show abstract

Section: A Sketching Algorithmmentioning

confidence: 99%

Section: Bmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Programmable Big Data Processing Framework to Reduce On-Chip Communications and Computations Towards Reducing Energy of the Processing

Medavaka¹

2019

International Journal of Advanced Research in Computer and Comm

View full text Add to dashboard Cite

show abstract

“…There has also been a tremendous amount of work done on wireless RF modules ranging from specialized research modules to commercial modules such as Bluetooth Smart (17.9mA RX,18.2mA TX, 1uA sleep). Still the relatively high amount of power required to transmit raw or even compressed data, makes it essential to perform local onboard processing [5]. The enterprise of this work is on the processor architecture that addresses the unique challenges and characteristics of biomedical applications.…”

Section: Introductionmentioning

confidence: 99%

Low-Power Manycore Accelerator for Personalized Biomedical Applications

Page

Attaran

Shea

et al. 2016

Proceedings of the 26th Edition on Great Lakes Symposium on VLSI

Self Cite

View full text Add to dashboard Cite

Wearable personal health monitoring systems can offer a cost effective solution for human healthcare. These systems must provide both highly accurate, secured and quick processing and delivery of vast amount of data. In addition, wearable biomedical devices are used in inpatient, outpatient, and at home e-Patient care that must constantly monitor the patient's biomedical and physiological signals 24/7. These biomedical applications require sampling and processing multiple streams of physiological signals with strict power and area footprint. The processing typically consists of feature extraction, data fusion, and classification stages that require a large number of digital signal processing and machine learning kernels. In response to these requirements, in this paper, a lowpower, domain-specific manycore accelerator named Power Efficient Nano Clusters (PENC) is proposed to map and execute the kernels of these applications. Experimental results show that the manycore is able to reduce energy consumption by up to 80% and 14% for DSP and machine learning kernels, respectively, when optimally parallelized. The performance of the proposed PENC manycore when acting as a coprocessor to an Intel Atom processor is compared with existing commercial off-the-shelf embedded processing platforms including Intel Atom, Xilinx Artix-7 FPGA, and NVIDIA TK1 ARM-A15 with GPU SoC. The results show that the PENC manycore architecture reduces the energy by as much as 10X while outperforming all off-the-shelf embedded processing platforms across all studied machine learning classifiers. CCS Concepts •Computer systems organization → Architectures; •Computing methodologies → Parallel computing methodologies; Machine learning algorithms; •Applied computing → Bioinformatics;

show abstract