FPGA Implementation of a Low-Power QRS Extractor

Self Cite

Convolutional Neural Networks (CNNs) are the state-of-the-art in computer vision for different purposes such as image and video classification, recommender systems and natural language processing. The connectivity pattern between CNNs neurons is inspired by the structure of the animal visual cortex. In order to allow the processing, they are realized with multiple parallel 2dimensional FIR filters that convolve the input signal with the learned feature maps. For this reason, a CNN implementation requires highly parallel computations that cannot be achieved using traditional general-purpose processors, which is why they benefit from a very significant speed-up when mapped and run on Field Programmable Gate Arrays (FPGAs). This is because FPGAs offer the capability to design full customizable hardware architectures, providing high flexibility and the availability of hundreds to thousands of on-chip Digital Signal Processing (DSP) blocks. This paper presents an FPGA implementation of a handwritten number recognition system based on CNN. The system has been characterized in terms of classification accuracy, area, speed, and power consumption. The neural network was implemented on a Xilinx XC7A100T FPGA, and it uses 29.69% of Slice LUTs, 4.42% of slice registers and 52.50% block RAMs. We designed the system using a 9-bit representation that allows for avoiding the use of DSP. For this reason, multipliers are implemented using LUTs. The proposed architecture can be easily scaled on different FPGA devices thank its regularity. CNN can reach a classification accuracy of 90%.

Section: Resultsmentioning

confidence: 99%

FPGA Implementation of Hand-written Number Recognition Based on CNN

Giardino

Matta

Silvestri

et al. 2019

Self Cite

“…Results show that the relationships between ECG features and model parameters present a non-linear trend and that the fitting neural networks can generalize them with an acceptable error. Considering the enormous spread of IoT devices [38], this type of system, as well as other complex systems useful for health care [39], will necessarily have to be integrated into low energy consumption portable embedded devices [40]-42] which will allow monitoring the quality of patients' health more efficiently.…”

Section: Discussionmentioning

confidence: 99%

Relationship between Mathematical Parameters of Modified Van der Pol Oscillator Model and ECG Morphological Features

Silvestri¹,

Acciarito²,

Khanal³

2019

Self Cite

The mathematical model describes the electrical and mechanical activity of the cardiac conduction system thought set of differential equations. By changing the value of parameters included in these equations, it is possible to change the amplitude and the period of ECG waves. Although this model is a powerful tool for modeling the electrical activity of the heart, its use is often limited to those familiar with the differential equations that describe the system. The purpose of this work is to provide a system that allows generating an ECG signal using Ryzhii model without knowing the details of differential equations. First, we provide the relationships between the ECG wave features and the model parameters; then we generalize them through fitting neural networks. Finally, putting in series fitting neural network and heart model, we provide a system that allows generating a synthetic signal by setting as input only the morphological ECG feature. We computed numerical simulation in Simulink environment and implemented the fitting neural networks in Matlab. Results show that non-linear trends characterize the correlation functions between ECG morphological features and model parameters and that the fitting neural networks can generalized this trend by providing the model parameters given in input the respectively ECG feature.

“…Exposing the used cooking oils to an UV radiation (peaked @ 400 nm) we obtained a fluorescence response with a maximum at 625 nm typical for the olive oil, see Fig. 4, developed using a hardware architecture [18].…”

Section: A Greenbag Binmentioning

confidence: 99%

The greenBag, the New Solution in Waste Separation

Sperandio¹,

Luca

Catini

2019

The greenBag is an innovative households' Used Cooking Oil (UCO) disposal solution aiming to allow the collection of this feedstock interesting European but also Chinese and US market. The possibility to collect UCO produced by households represents the possibilities to increase the amount of biodiesel, respecting the ILUC directive, of the 60%. So far, there are no concrete solutions able to properly respond to this need, due to their huge prizes and due to the low performances, they provide. greenBag is a solution coming from an intensive research activity completed by an infield validation. It has been designed to reply to all the requirements coming from families and from the waste management companies, the two actors involved in UCO disposal. Indeed, two different types of users in general, and in particular the UCO collection, characterize waste collection solutions: families-that want simple solutions, easy to use and able to reward them; waste collection companies-that want cheap solutions, able to trace waste disposals and able to assure them a high quality of wastes. The currently available market solutions are all focused on waste companies and the current innovative ideas designed by Universities and Startups focus their attention only on one of the two involved actors. greenBag is currently the sole solution that aims to satisfy both actors, exploiting the disposal traceability not only to know how many waste people produces but also to quantify their efforts and making them aware about the overall disposal process.