Perspectives and Challenges in Doctoral Research—Selected Papers from the 10th Edition of the Scientific Conference of the Doctoral Schools of “Dunărea de Jos” University of Galati (SCDS-UDJG)

A special issue of Inventions (ISSN 2411-5134).

Deadline for manuscript submissions: closed (31 December 2022) | Viewed by 36612

Printed Edition Available!
A printed edition of this Special Issue is available here.

Special Issue Editors


E-Mail Website
Guest Editor
Department of Applied Mechanics, "Dunarea de Jos" University of Galati, Galați, Romania
Interests: renewable energy; marine engineering; offshore technologies
Special Issues, Collections and Topics in MDPI journals

Special Issue Information

Dear Colleagues,

We are inviting submissions to a Special Issue of the Invention journal for participation in the 10th edition of the Scientific Conference of the Doctoral Schools from “Dunărea de Jos” University of Galaţi. The objective of the 2022 Conference is to bring together perspectives and challenges in doctoral research to a common forum. The aim of the conference is to provide a platform to the doctoral researchers to meet and share state of the art developments in their field. On this occasion, our institution targets to promote excellence in research, to set up partnerships and collaborative relationships through the exchange of knowledge and expertise. As in the previous editions, the conference invites oral and poster presentations in sections related to the main domains of the doctoral research at UDJG. Workshops, exhibition stands, and social activities are also included in the programme, all with the aim of developing and improving the network of the doctoral schools.

All the papers presented in the conference and accepted for publication in this Special Issue will benefit of a 50% discount for the APC.

Prof. Dr. Eugen Rusu
Prof. Dr. Gabriela Rapeanu
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Inventions is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 1800 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Keywords

  • advanced research in mechanical and industrial engineering
  • progress in food science and bio-resources engineering
  • advances in engineering and management in agriculture and rural development
  • advanced research in electrical/electronic engineering, system engineering and information technologies
  • future of eco-nanotechnologies, functional materials and coatings
  • chemistry
  • electrochemistry in life sciences

Published Papers (14 papers)

Order results
Result details
Select all
Export citation of selected articles as:

Editorial

Jump to: Research

3 pages, 210 KiB  
Editorial
Editorial Note for the Special Issue: Perspectives and Challenges in Doctoral Research—Selected Papers from the 10th Edition of the Scientific Conference of the Doctoral Schools from the “Dunărea de Jos”
by Eugen Rusu and Gabriela Rapeanu
Inventions 2023, 8(3), 73; https://doi.org/10.3390/inventions8030073 - 17 May 2023
Viewed by 1030
Abstract
This editorial note is dedicated to the 10th Scientific Conference which was held on June 2022 in Galati, Romania, and was organized by the Council of Doctoral Schools of the “Dunărea de Jos” University of Galati (SCDS-UDJG) [...] Full article

Research

Jump to: Editorial

13 pages, 1598 KiB  
Article
Optimization of Extraction Parameters of Anthocyanin Compounds and Antioxidant Properties from Red Grape (Băbească neagră) Peels
by Daniela Serea, Oana Emilia Constantin, Georgiana Horincar, Nicoleta Stănciuc, Iuliana Aprodu, Gabriela Elena Bahrim and Gabriela Râpeanu
Inventions 2023, 8(2), 59; https://doi.org/10.3390/inventions8020059 - 27 Mar 2023
Cited by 3 | Viewed by 1662
Abstract
Using a Central Composite Design, the extraction of bioactive compounds from red grape Băbească neagră peels was optimized by applying a conventional solvent extraction. On the anthocyanin content, total phenolic content, and antioxidant activity (using the DPPH method), the effects of extraction factors, [...] Read more.
Using a Central Composite Design, the extraction of bioactive compounds from red grape Băbească neagră peels was optimized by applying a conventional solvent extraction. On the anthocyanin content, total phenolic content, and antioxidant activity (using the DPPH method), the effects of extraction factors, including ethanol and citric acid concentrations, extraction temperature, and duration, were investigated. For each of the investigated parameters, a quadratic model was suggested. The maximum and minimum variables investigated in the coded form of the experimental plan are the concentrations of citric acid (0.10–2.64%), ethanol (38.06–96.93%), operating temperature (13.06–71.90 °C), and extraction time (11.36–78.63 min). The optimal mixture for recovering the most significant amount of polyphenol content and antioxidant activity was 85% ethanol, 0.85% citric acid, 52.14 min, and 57 °C. Based on the experimental approach, the anthocyanin content ranged from 1.71 to 2.74 mg C3G/g DW, the total phenolic content ranged from 24.67 to 43.97 mg/g, and the antioxidant activity ranged from 15.95 to 20.98 mM TE/g DW. Overall, it should be stressed that establishing operating factors to maximize model responses can improve the extraction process and the obtaining of red grape peel value-added extracts for creating functional food products. Full article
Show Figures

Figure 1

15 pages, 3408 KiB  
Article
Machine Learning Systems Detecting Illicit Drugs Based on Their ATR-FTIR Spectra
by Iulia-Florentina Darie, Stefan Razvan Anton and Mirela Praisler
Inventions 2023, 8(2), 56; https://doi.org/10.3390/inventions8020056 - 13 Mar 2023
Cited by 1 | Viewed by 2514
Abstract
We present a comparative study aiming to determine the most efficient multivariate model screening for the main drugs of abuse based on their ATR-FTIR spectra. A preliminary statistical analysis of selected spectra data extracted from the public SWGDRUG IR Library was first performed. [...] Read more.
We present a comparative study aiming to determine the most efficient multivariate model screening for the main drugs of abuse based on their ATR-FTIR spectra. A preliminary statistical analysis of selected spectra data extracted from the public SWGDRUG IR Library was first performed. The results corroborated those of an exploratory analysis that was based on several dimensionality reduction methods, i.e., Principal Component Analysis (PCA), Independent Component Analysis (ICA), and autoencoders. Then, several machine learning methods, i.e., Support Vector Machines (SVM), eXtreme Gradient Boosting (XGB), Random Forest, Gradient Boosting, and K-Nearest Neighbors (KNN), were used to assign the drug class membership. In order to account for the stochastic nature of these machine learning methods, both models were evaluated 10 times on a randomly distributed subset of the whole SWGDRUG IR Library, and the results were compared in detail. Finally, their performance in assigning the class identity of three classes of drugs of abuse, i.e., hallucinogenic (2C-x, DOx, and NBOMe) amphetamines, cannabinoids, and opioids, were compared based on confusion matrices and various classification parameters, such as balanced accuracy, sensitivity, and specificity. The advantages of each of the illicit drug-detecting systems and their potential as forensic screening tools used in field scenarios are also discussed. Full article
Show Figures

Figure 1

16 pages, 3830 KiB  
Article
An Evaluation of the Wind Energy along the Romanian Black Sea Coast
by Laura-Ionela Nedelcu, Viorel-Mihai Tanase and Eugen Rusu
Inventions 2023, 8(1), 48; https://doi.org/10.3390/inventions8010048 - 13 Feb 2023
Cited by 4 | Viewed by 1860
Abstract
The present study aims to outline a general overview of the wind energy potential along the Romanian coast of the Black Sea, using the weather data provided by the Maritime Hydrographic Directorate covering a 13-year time interval (2009–2021). The data obtained from seven [...] Read more.
The present study aims to outline a general overview of the wind energy potential along the Romanian coast of the Black Sea, using the weather data provided by the Maritime Hydrographic Directorate covering a 13-year time interval (2009–2021). The data obtained from seven automatic weather coastal stations distributed along the Romanian perimeter were used to evaluate the wind regime, highlighting the Black Sea’s complex marine environment. The analysis based on the evaluation of the wind parameters per each station registered on the total period revealed that the overall wind characteristics are similar, resulting in no significant variations depending on the station’s location. Moreover, the climatic picture of the Black Sea can be interpreted as two seasons, winter and summer, a conclusion based on the analysis made of the seasonal and monthly variation of the wind aspects. Subsequently, the outcomes obtained in this research imply that the Romanian Black Sea coast has the potential to be a good location for wind energy development due to the strong winds that blow in the region. Full article
Show Figures

Figure 1

16 pages, 8355 KiB  
Article
The Influence of Different Type Materials of Grit Blasting on the Corrosion Resistance of S235JR Carbon Steel
by Nicoleta Bogatu, Alina Crina Muresan, Laurentiu Mardare, Viorica Ghisman, Anca Ravoiu, Floricel Maricel Dima and Daniela Laura Buruiana
Inventions 2023, 8(1), 39; https://doi.org/10.3390/inventions8010039 - 2 Feb 2023
Cited by 1 | Viewed by 2751
Abstract
The aim of this paper is to evaluate the corrosion rate expressed in material loss per unit of time and the surface properties of carbon steel type S235JR blasted with different types of materials (quartz, alumina, and red garnet with a particle size [...] Read more.
The aim of this paper is to evaluate the corrosion rate expressed in material loss per unit of time and the surface properties of carbon steel type S235JR blasted with different types of materials (quartz, alumina, and red garnet with a particle size between 60 and 80 mesh (0.25–0.60 mm)). The estimation of corrosion rate was determined by electrochemical methods, such as open circuit potential (OCP), polarization resistance (Rp), corrosion rate (Vcorr), and gravimetric method by immersing the samples in 3.5% NaCl solution for a period of 336 h. All surfaces were characterized before and after corrosion tests using ex-situ characterizations, such as optical microscopy and roughness analysis. The results indicate that S235JR non-sandblasted exhibited higher polarization resistance, the lowest corrosion rate, and the lowest roughness values. While for the S235JR sandblasted groups, reduced corrosion resistance and increasing roughness values were noted. From the sandblasted groups, the lowest corrosion resistance and the highest value of roughness are attributed to the S235JR surface sandblasted with quartz. The S235JR surface sandblasted with quartz shows a decrease in corrosion resistance approximately two times lower than the non-sandblasted surface and an increasing of roughness approximately six times greater than the non-sandblasted surface. Full article
Show Figures

Figure 1

16 pages, 2256 KiB  
Article
Study on the Design Stage from a Dimensional and Energetic Point of View for a Marine Technical Water Generator Suitable for a Medium Size Container Ship
by Catalin Faitar and Eugen Rusu
Inventions 2023, 8(1), 22; https://doi.org/10.3390/inventions8010022 - 12 Jan 2023
Cited by 1 | Viewed by 1511
Abstract
The purpose of this study is to provide an overview of the development of the modern low-speed marine two-stroke diesel engine from the point of view of the technical water cooling plant, taking into account and starting from the market requirements for power [...] Read more.
The purpose of this study is to provide an overview of the development of the modern low-speed marine two-stroke diesel engine from the point of view of the technical water cooling plant, taking into account and starting from the market requirements for power and speed, with information and design options relevant to the entire shipping industry. Thus, through the ideas of this project, we analyze notions and relevant aspects of systems related to marine slow turning engines, including the basic thermodynamic structure of the technical water system designed for a marine engine. This study also presents the design criteria that define the size and design concept of the engine structure components, with a focus on the technical water cooling installation. The concepts for the main engine hot parts served by the technical water cooling installation play a vital role in the marine technical water generator. The choices the engine designer must make regarding basic auxiliary systems, such as fuel injection and exhaust valve actuation, are important factors to keep in mind when installing a technical water generator onboard a ship. The automation and control systems that govern the modern electronic engine, which drive the supply pump of the technical water cooling system can provide a simplified view of the engine development process. In order to point out the contributions of this study, it is important to focus on the calculations used to determine the main parameters of a technical water generator especially designed for a midsized container ship. Full article
Show Figures

Figure 1

12 pages, 2441 KiB  
Article
Optimization of the Parameters Influencing the Antioxidant Activity and Concentration of Anthocyanins Extracted from Red Onion Skins Using a Central Composite Design
by Florina Stoica, Oana Emilia Constantin, Nicoleta Stănciuc, Iuliana Aprodu, Gabriela Elena Bahrim and Gabriela Râpeanu
Inventions 2022, 7(4), 89; https://doi.org/10.3390/inventions7040089 - 3 Oct 2022
Cited by 2 | Viewed by 2000
Abstract
This study aimed to extract bioactives from red onion skins for use as edible colorants that are both natural and functional. The extraction of bioactive chemicals from red onion skins using a conventional solvent extraction was optimized using a Central Composite Design (CCD). [...] Read more.
This study aimed to extract bioactives from red onion skins for use as edible colorants that are both natural and functional. The extraction of bioactive chemicals from red onion skins using a conventional solvent extraction was optimized using a Central Composite Design (CCD). The influence of extraction parameters, such as ethanol and citric acid concentrations, extraction temperature, and time, on anthocyanin content and antioxidant activity (DPPH method) was studied. A quadratic model was suggested for all of the parameters examined and employed. Citric acid concentration (0.05–2.64%), ethanol concentration (6.36–73.63%), operation temperature (16.47–58.52 °C), and extraction duration (10–234.54 min) were the variables studied in the coded form of the experimental plan. The best conditions for maximum anthocyanins and antioxidant activity recovery were: 60% ethanol, 0.87% citric acid, 179.99 min, and 25 °C. The anthocyanins concentration varied from 0.45 to 1.43 mg C3G/g DW, while the antioxidant activity varied from 24.29 to 37.20 mM TE/g DW, according to the experimental design. Overall, it should be emphasized that the extraction process can be enhanced by settling the operating factors to maximize the model responses. The current findings demonstrate that extracts from red onion skins would be useful in developing functional food products. Full article
Show Figures

Figure 1

16 pages, 3927 KiB  
Article
Event-Based PID Control of a Flexible Manufacturing Process
by Octavian Duca, Eugenia Minca, Adrian Filipescu, Daniela Cernega, Razvan Solea and Claudiu Bidica
Inventions 2022, 7(4), 86; https://doi.org/10.3390/inventions7040086 - 26 Sep 2022
Cited by 3 | Viewed by 1921
Abstract
In most cases, the system control is made in a sampled manner, measuring the controlled value at a predefined frequency given by the sampling time. However, not all processes provide relevant information at regular intervals, especially in manufacturing. To reduce the costs and [...] Read more.
In most cases, the system control is made in a sampled manner, measuring the controlled value at a predefined frequency given by the sampling time. However, not all processes provide relevant information at regular intervals, especially in manufacturing. To reduce the costs and complexity of systems, event-based measuring is necessary. To control this kind of process, an event-based controller is needed. This poses some challenges, especially between the event-triggered measurement, as the process runs in an open loop. In the literature, most event-based controllers are based on the comparison of the error value with a predefined value and activate the controller if this value is crossed. However, in this type of controller, the measured value is measured at a predefine interval and is not suited for most event-based processes. In manufacturing systems, the most usual event-based process is represented by the conveyor transportation system. In this process, the product position is measured only in key locations on the conveyor. For the optimal operation of a flexible manufacturing system, the presence of a product in a key location at predetermined intervals is necessary. For this purpose, this article presents an event-based PID controller implemented on a conveyor transportation system. Full article
Show Figures

Figure 1

6 pages, 658 KiB  
Article
Sensitivity Analysis of Artificial Neural Networks Identifying JWH Synthetic Cannabinoids Built with Alternative Training Strategies and Methods
by Catalina Mercedes Burlacu, Adrian Constantin Burlacu and Mirela Praisler
Inventions 2022, 7(3), 82; https://doi.org/10.3390/inventions7030082 - 13 Sep 2022
Cited by 4 | Viewed by 1971
Abstract
This paper presents the alternative training strategies we tested for an Artificial Neural Network (ANN) designed to detect JWH synthetic cannabinoids. In order to increase the model performance in terms of output sensitivity, we used the Neural Designer data science and machine learning [...] Read more.
This paper presents the alternative training strategies we tested for an Artificial Neural Network (ANN) designed to detect JWH synthetic cannabinoids. In order to increase the model performance in terms of output sensitivity, we used the Neural Designer data science and machine learning platform combined with the programming language Python. We performed a comparative analysis of several optimization algorithms, error parameters and regularization methods. Finally, we performed a new goodness-of-fit analysis between the testing samples in the data set and the corresponding ANN outputs in order to investigate their sensitivity. The effectiveness of the new methods combined with the optimization algorithms is discussed. Full article
Show Figures

Figure 1

17 pages, 5643 KiB  
Article
Wind Variation near the Black Sea Coastal Areas Reflected by the ERA5 Dataset
by Victoria Yildirir, Eugen Rusu and Florin Onea
Inventions 2022, 7(3), 57; https://doi.org/10.3390/inventions7030057 - 7 Jul 2022
Cited by 2 | Viewed by 2155
Abstract
In the context of the European Green Deal implementation, it is expected that there will be an increase in number of the wind farms located near the coastal areas in order to support this initiative. The Black Sea represents an important source of [...] Read more.
In the context of the European Green Deal implementation, it is expected that there will be an increase in number of the wind farms located near the coastal areas in order to support this initiative. The Black Sea represents an important source of wind energy, and as a consequence, in the present work the regional wind resources (onshore and offshore) are evaluated by considering a total of 20 years of ERA5 wind data covering the 20-year time interval from January 2002 to December 2021. From a general perspective, it is clear that the offshore areas (100 km from the shoreline) are defined by much higher wind speed values than in the onshore, reaching an average of 8.75 m/s for the points located on the western sector. During the winter, these values can go up to 8.75 m/s, with the mention that the northern sectors from Ukraine and Russia may easily exceed 8 m/s. In terms of the wind turbines’ selection, for the offshore areas defined by consistent wind resources, generators will be considered that are defined by a rated wind speed of 11 m/s. Finally, we can mention that a theoretical offshore wind turbine of 20 MW can reach a capacity factor located between 20.9 and 48.3%, while a maximum annual electricity production of 84.6 GWh may be obtained from the sites located near the Romanian and Ukrainian sectors, respectively. Full article
Show Figures

Figure 1

20 pages, 14164 KiB  
Article
Mobile Visual Servoing Based Control of a Complex Autonomous System Assisting a Manufacturing Technology on a Mechatronics Line
by Georgian Simion, Adrian Filipescu, Dan Ionescu, Răzvan Șolea, Daniela Cernega, Eugenia Mincă and Adriana Filipescu
Inventions 2022, 7(3), 47; https://doi.org/10.3390/inventions7030047 - 22 Jun 2022
Cited by 3 | Viewed by 2009
Abstract
The main contribution of this paper is the modeling and control for a complex autonomous system (CAS). It is equipped with a visual sensor to operate precision positioning in a technology executed on a laboratory mechatronics line. The technology allows the retrieval of [...] Read more.
The main contribution of this paper is the modeling and control for a complex autonomous system (CAS). It is equipped with a visual sensor to operate precision positioning in a technology executed on a laboratory mechatronics line. The technology allows the retrieval of workpieces which do not completely pass the quality test. Another objective of this paper is the implementation of an assisting technology for a laboratory processing/reprocessing mechatronics line (P/RML) containing four workstations, assisted by the following components: a complex autonomous system that consists of an autonomous robotic system (ARS), a wheeled mobile robot (WMR) PeopleBot, a robotic manipulator (RM) Cyton 1500 with seven degrees of freedom (7 DOF), and a mobile visual servoing system (MVS) with a Logitech camera as visual sensor used in the process of picking, transporting and placing the workpieces. The purpose of the MVS is to increase the precision of the RM by utilizing the look and move principle, since the initial and final positions of the CAS can slightly deviate from their trajectory, thus increasing the possibility of errors to appear during the process of catching and releasing the pieces. If the processed piece did not pass the quality test and has been rendered as defective, it is retrieved from the last station of the P/RML and transported to the first station for reprocessing. The control of the WMR is done using the trajectory-tracking sliding-mode control (TTSMC). The RM control is based on inverse kinematics model, and the MVS control is implemented with the image moments method. Full article
Show Figures

Figure 1

28 pages, 21599 KiB  
Article
Communication and Control of an Assembly, Disassembly and Repair Flexible Manufacturing Technology on a Mechatronics Line Assisted by an Autonomous Robotic System
by Dan Ionescu, Adrian Filipescu, Georgian Simion, Eugenia Mincă, Daniela Cernega, Răzvan Șolea and Adriana Filipescu
Inventions 2022, 7(2), 43; https://doi.org/10.3390/inventions7020043 - 15 Jun 2022
Cited by 7 | Viewed by 6525
Abstract
This paper aims to describe modeling and control in what concerns advanced manufacturing technology running on a flexible assembly, disassembly and repair on a mechatronic line (A/D/RML) assisted by an Autonomous Robotic System (ARS), two robotic manipulators (RM) and visual servoing system (VSS). [...] Read more.
This paper aims to describe modeling and control in what concerns advanced manufacturing technology running on a flexible assembly, disassembly and repair on a mechatronic line (A/D/RML) assisted by an Autonomous Robotic System (ARS), two robotic manipulators (RM) and visual servoing system (VSS). The A/D/RML consists of a six workstations (WS) mechatronics line (ML) connected to a flexible cell (FC) equipped with a 6-DOF ABB industrial robotic manipulator (IRM) and an ARS used for manipulation and transport. A hybrid communication and control based on programmable logic controller (PLC) architecture is used, which consists of two interconnected systems that feature both distributed and centralized topology, with specific tasks for all the manufacturing stages. Profinet communication link is used to interconnect and control FC and A/D/RML. The paper also discusses how to synchronize data between different field equipment used in the industry and the control systems. Synchronization signals between the master PLC and ARS is performed by means of Modbus TCP protocol and OPC UA. The structure of the ARS consists of a wheeled mobile robot (WMR) with two driving wheels and one free wheel (2DW/1FW) equipped with a 7-DOF RM. Trajectory tracking sliding-mode control (TTSMC) is used to control WMR. The end effector of the ARS RM is equipped with a mobile eye-in-hand VSS technology for the precise positioning of RM to pick and place the workparts in the desired location. Technology operates synchronously with signals from sensors and from the VSS HD camera. If the workpiece does not pass the quality test, the process handles it by transporting back from the end storage unit to the flexible cell where it will be considered for reprocessing, repair or disassembling with the recovery of the dismantled parts. The recovered or replaced components are taken over by the ARS from disassembling location and transported back to the dedicated storage warehouses to be reused in the further assembly processes. Full article
Show Figures

Figure 1

11 pages, 1987 KiB  
Article
Image Moment-Based Features for Mass Detection in Breast US Images via Machine Learning and Neural Network Classification Models
by Iulia-Nela Anghelache Nastase, Simona Moldovanu and Luminita Moraru
Inventions 2022, 7(2), 42; https://doi.org/10.3390/inventions7020042 - 15 Jun 2022
Cited by 4 | Viewed by 2056
Abstract
Differentiating between malignant and benign masses using machine learning in the recognition of breast ultrasound (BUS) images is a technique with good accuracy and precision, which helps doctors make a correct diagnosis. The method proposed in this paper integrates Hu’s moments in the [...] Read more.
Differentiating between malignant and benign masses using machine learning in the recognition of breast ultrasound (BUS) images is a technique with good accuracy and precision, which helps doctors make a correct diagnosis. The method proposed in this paper integrates Hu’s moments in the analysis of the breast tumor. The extracted features feed a k-nearest neighbor (k-NN) classifier and a radial basis function neural network (RBFNN) to classify breast tumors into benign and malignant. The raw images and the tumor masks provided as ground-truth images belong to the public digital BUS images database. Certain metrics such as accuracy, sensitivity, precision, and F1-score were used to evaluate the segmentation results and to select Hu’s moments showing the best capacity to discriminate between malignant and benign breast tissues in BUS images. Regarding the selection of Hu’s moments, the k-NN classifier reached 85% accuracy for moment M1 and 80% for moment M5 whilst RBFNN reached an accuracy of 76% for M1. The proposed method might be used to assist the clinical diagnosis of breast cancer identification by providing a good combination between segmentation and Hu’s moments. Full article
Show Figures

Figure 1

13 pages, 4574 KiB  
Article
A Fingerprint Matching Algorithm Using the Combination of Edge Features and Convolution Neural Networks
by Andreea-Monica Dincă Lăzărescu, Simona Moldovanu and Luminita Moraru
Inventions 2022, 7(2), 39; https://doi.org/10.3390/inventions7020039 - 27 May 2022
Cited by 8 | Viewed by 4812
Abstract
This study presents an algorithm for fingerprint classification using a CNN (convolutional neural network) model and making use of full images belonging to four digital databases. The main challenge that we face in fingerprint classification is dealing with the low quality of fingerprints, [...] Read more.
This study presents an algorithm for fingerprint classification using a CNN (convolutional neural network) model and making use of full images belonging to four digital databases. The main challenge that we face in fingerprint classification is dealing with the low quality of fingerprints, which can impede the identification process. To overcome these restrictions, the proposed model consists of the following steps: a preprocessing stage which deals with edge enhancement operations, data resizing, data augmentation, and finally a post-processing stage devoted to classification tasks. Primarily, the fingerprint images are enhanced using Prewitt and Laplacian of Gaussian filters. This investigation used the fingerprint verification competition with four databases (FVC2004, DB1, DB2, DB3, and DB4) which contain 240 real fingerprint images and 80 synthetic fingerprint images. The real images were collected using various sensors. The innovation of the model is in the manner in which the number of epochs is selected, which improves the performance of the classification. The number of epochs is defined as a hyper-parameter which can influence the performance of the deep learning model. The number of epochs was set to 10, 20, 30, and 50 in order to keep the training time at an acceptable value of 1.8 s/epoch, on average. Our results indicate the overfitting of the model starting with the seventh epoch. The accuracy varies from 67.6% to 98.7% for the validation set, and between 70.2% and 75.6% for the test set. The proposed method achieved a very good performance compared to the traditional hand-crafted features despite the fact that it used raw data and it does not perform any handcrafted feature extraction operations. Full article
Show Figures

Figure 1

Back to TopTop