Showing papers by "Consolatina Liguori published in 2019"

Automatic Detection of Melanoma with Yolo Deep Convolutional Neural Networks

Abstract: In the past three years, deep convolutional neural networks (DCNNs) have achieved promising results in detecting skin cancer. However, improving the accuracy and efficiency of the automatic detection of melanoma is still urgent due to the visual similarity of benign and malignant dermoscopic images. There is also a need for fast and computationally effective systems for mobile applications targeting caregivers and homes. This paper presents the You Only Look Once (Yolo) algorithms, which are based on DCNNs applied to the detection of melanoma. The Yolo algorithms comprise YoloV1, YoloV2, and YoloV3, whose methodology first resets the input image size and then divides the image into several cells. According to the position of the detected object in the cell, the network will try to predict the bounding box of the object and the class confidence score. Our test results indicate that the mean average precision (mAP) of Yolo can exceed 0.82 with a training set of only 200 images, proving that this method has great advantages for detecting melanoma in lightweight system applications.

Data Fusion for Timber Bundle Volume Measurement

Abstract: In many industrial fields, the direct measurement of a specific parameter of interest can be too costly to be feasible. Using data from other parameters about the process, a data driven approach can instead be used for a cost-efficient estimation of the parameter of interest. For the Swedish forest industry, the measurement of the timber bundles volume sets the price for the customers. Volume is in this case an expensive parameter to measure, where manual measurements have been used as reference to a remote imaging estimation system involving operators. In this paper a data driven approach is presented for the estimation of timber bundles volume, using correlated features such as weight, date of cut, storage time etc., to demonstrate a costefficient software estimation system. The proposed approach, based on neural networks, shows comparable results to the existing remote estimation method when using a few features and even better performance when adding feature information from the timber harvesters.

Smart meters and water leakage detection: a preliminary study

Abstract: The paper deals with water leakage beyond the meter at household level. The automatic leakage detection is emerging as priority task for scientists and technicians. Each possible solution must include the use of smart meters for the continuous survey of domestic water consumption. Water smart meters aren’t widespread yet, and the few either are prototypes or exhibit not adequate battery and communication performance. The authors, participants to the EMPIR 17IND13 Metrowamet project, propose an embedded system that can work with different water sensors, whose characteristics have been suitably designed for respecting stringent requirements of energy efficiency, computational burden and wireless transmission capability.

A Statistical Approach for Improving the Accuracy of Dry Friction Coefficient Measurement

Abstract: Tribological tests are adopted to estimate the kinetic friction coefficient of the material of interest according to the ASTM standard. Typically, for the measurement process, several replicas as well as a postprocessing data treatment are necessary to take into account the observed casual variability of the measurand. This paper describes a novel statistical approach aiming to highlight the most significant time intervals during each test replica. The experimental activity has been carried out with reference to the ball-on-flat configuration of the tribometer in order to verify the suggested procedure in terms of both accuracy and reliability. Numerical simulation and comparison of the measurement results with the output of a different outlier detection algorithm are also included to show the improvement allowed by the authors’ proposal about the dry friction coefficient measurement.

Smart Devices and Services for Smart City

Abstract: Citizen quality of life can be improved through facilities and services that have been thought to ease citizen interaction with municipal authorities, offices and structures. All technologies and devices, used for developing these facilities, are the pillars of the Smart City idea: a City that adapts itself, at least in part, to citizens’ needs. Advanced Metering Infrastructure (AMI) could become the backbone of all the smart city projects. Other public services can be loaded on AMI’s to be smart and thus helping to find the affordability of investments. The paper deals with this topic by describing devices and results of a pilot project, which has been carried out in an Italian middle city (Salerno), to experience the use of RF 169MHz wM-bus based AMI. Experimental results regarding a set of about 2500 installed devices for gas and water metering, car parking management and elder teleassistance, will be reported in detail to show convenience and problems of this approach.

Bridge dynamic analysis for accurate control and monitoring: Feasibility study

Abstract: Most of engineering infrastructure including bridges impacts significantly civilians life quality and safety. Despite carrying their own weight and live load during operation, structure members are also seriously affected by disaster and environment. This research paper proposed an investigation in the changes of vibration frequency and the main girder(road) deflection. The outcome of the dynamic analysis that considered a vehicle traveling on the bridge, demonstrate that the vibration characteristics and the deflection response changed as damage progressed. It was discovered to be possible to identify the conditions leading to advanced deformations and collapse of the bridge. The paper finally proposes an sensor network able to identify the earliest damage which can occurs on the bridge.

Enhanced Methods for Extracting Characteristic Features from ECG

Abstract: QRS estimation is an assessment allowing to understand manifold indications related to heart operating mode, hidden functionalities and abnormalities due to diverse reasons that sometimes are not apparently correlated to circulatory reasons. The paper illustrates an enhanced version of Pan and Tompkin’s QRS detector for ECG signal processing. Two patients’ ECG signals are under exam for trying to recover their heart conditions. An preliminary insight ECG for cyber-physical system (CPS), then for IoT is traced-out.

A low-cost device for beer color measurement

Abstract: Color is an important quality parameter in the beer production. It is usually measured by means of a spectrophotometric method that has been adopted as an international standard by the main brewers associations in the world. The instruments based on this method are always present in the quality laboratory of the industrial breweries, but their cost is often not affordable for the microbreweries that need inexpensive solutions to evaluate the quality parameters of their processes and products. In this work, a low-cost spectrophotometer for the measurement of the color of beer, based on the European Brewery Convention (EBC) international method, is proposed.

Quality control of hazelnuts by means of NMR measurements

Abstract: Hazelnuts are widely used in the confectionary industry for their flavour and taste. In order to guarantee a suitable level of quality, several standards have been defined by international organizations and producing countries. They define the minimum quality requirements of the fruits in terms of dimension, aspect, level of moisture, hidden defects. In this framework, our proposal is related to the set-up of an in-line classification system, based on low field NMR, able to detect the hidden defects of the hazelnuts. The proposed classification procedure is based on the CPMG sequence and the analysis of the transverse relaxation decay. The procedure includes three steps in which different features are detected: (i) moisture content, (ii) kernel development and (iii) healthy detection (presence of mould). Experimental results showed a very good capability to correctly detect the hidden defects, obtaining a sensitivity of 95%, with a specificity (correct detection of the healthy hazelnuts) around the 80%.

Accurate measurement of Kinetic Friction Coefficient by using two types of tribometer

Abstract: Tribological tests are adopted to estimate the kinetic friction coefficient of the material of interest according to the ASTM standard. Typically, for the measurement process, several replicas as well as a post-processing data treatment are necessary to take into account the observed casual variability of the measurand. The paper describes a statistical approach for both Time and Frequency Domain analysis, aiming to highlight the most significant time intervals during each test replica. Experimental activity has been carried out by performing multiple tests through two different tribometers (ball-on-flat and ball-on-disk configurations) in order to verify the improvement allowed by the authors’ proposal about the dry friction coefficient measurement in terms of both accuracy and reliability.

Modeling Stresses in Plasmonic Nanosensors

Abstract: Sensors are useful in many fields and in particular in environmental monitoring. Sea pollution detection is one of the special areas of environment matrices to be put under control. Given the fact, sensors, located in sea surface and/or deepness, are subject to temperature and mechanical stresses. This latter is a constrained burden in designing sensing systems for this scope. The paper illustrates the design of sensing system including a case hosting a nanosensor in order to allow sea water to pass through. A mechanical stress analysis has been performed.