Publications Archives - Page 6 of 8

04.08.2022.27.10.2022.

IEEE COINS 2022: HPC and Deep Learning for Computer Vision in Smart Farms

Researchers from EuroCC Montenegro presented two papers at the IEEE International Conference on Omni-Layer Intelligent Systems (COINS). IEEE COINS (link) is the right place to be. IEEE COINS brings together experts in Digital Transformation (from AI and IoT to Cloud, Blockchain, Cybersecurity, and Robotics) from around the globe. IEEE COINS includes a multi-disciplinary program from technical research papers, to panels, workshops, and tutorials on the latest technology developments and innovations addressing all important aspects of the IoT & AI ecosystem. The conference took place 1-3 August in Barcelona.

This paper was a result of the collaboration on FF4EuroHPC application experiment project called AIMHiGH that focuses on computer vision and the use of HPC to develop object detection prediction models for the use in smart agriculture, more specifically in the poultry sector. The title of the paper is “Developing Object Detection Models for Camera Applications in Smart Poultry Farms”.

ABSTRACT – This paper proposes the use of high-performance computing and deep learning to create prediction models that can be deployed as a part of smart agriculture solutions in the poultry sector. The idea is to create object detection models that can be ported onto edge devices equipped with camera sensors for the use in Internet of Things systems for poultry farms. The object detection prediction models could be used to create smart camera sensors that could evolve into sensors for counting chickens or detecting dead ones. Such camera sensor kits could become a part of digital poultry farm management systems in shortly. The paper discusses the approach to the development and selection of machine learning and computational tools needed for this process. Initial results, based on the use of Faster R-CNN network and high-performance computing are presented together with the metrics used in the evaluation process. The achieved accuracy is satisfactory and allows for easy counting of chickens. More experimentation is needed with network model selection and training configurations to increase the accuracy and make the prediction useful for developing a dead chicken detector. (link)

03.08.2022.27.10.2022.

IEEE COINS 2022: Detecting Pneumonia with TensorFlow and CNNs

Click on image to open the conference website

The paper called “Detecting Pneumonia with TensorFlow and Convolutional Neural Networks”, authored by D. Babic, I. Jovovic, T: Popovic, S. Cakic and L. Filipovic discussed the use of deep learning and HPC to create prediction models aimed at detecting pneumonia in chest x-ray images.

ABSTRACT – Artificial intelligence is getting more and more involved in our everyday life as a result of enormous amounts of data available for feeding the machine and deep learning algorithms. Deep learning introduced new dimensions and possibilities of applications in medical science. With COVID-19 outbreak in 2020 at global level, the health systems of many countries were overwhelmed. With many patients infected, health system is pressured to correctly diagnose patient’s state of illness. In a lot of occasions, it was almost impossible to correctly diagnose many COVID-19 positive patients that have pneumonia due to many outbreaks in many areas. The intelligent system that could detect pneumonia with certainty could help in easing the pressure on the health system and make doctors focus on more severely ill patients. This paper describes development of pneumonia detection model using TensorFlow to processes the chest X-ray images to determine whether the patient has pneumonia. The model is based on deep learning algorithm supported through convolutional neural network. The model presented in this paper has achieved rather high accuracy (over 95%) in analyzing X-Ray images and could be used to speed up decision process in healthcare. (link)

06.07.2022.27.10.2022.

BSc Thesis: The use of ML to assess plant health

Ms Almira Suljovic, a student of the Faculty of Applied Sciences, defended her BSc Thesis in Electrical Engineering and Computer Science. The topic of the thesis work was the use of machine learning to detect diseased plants by processing images of the leaves. The work included creation of a prediction model as well as the integration of the model into a mobile application for the use by farmers. She has done her thesis work under the supervision of prof. Tomo Popovic, PhD, and mr Stevan Cakic, MSc.

BSc Thesis – the use of ML to detect plant disease

ABSTRACT – Early detection and prevention are one of the biggest difficulties in the field of agriculture. Late detection of plant diseases or the use of wrong pesticides often leads to crop damage, reducing food quality. As the leaf is the best indicator of whether a plant is healthy or not, we can construct prediction models using machine learning to identify leaf status in the shortest possible time, thus preventing or reducing losses. This thesis shows how to determine leaf health status using the Detectron2 software library and the faster R-CNN neural network. The model was trained using a dataset with 6407 images. The initial dataset was extended using the RoboFlow tool. Google Colab, an environment for the development of cloud computing and machine learning, was used for testing and implementation. The practical application of the machine learning model was realized using an application developed using the Flutter platform.

27.06.2022.27.10.2022.

Master Thesis: Practical Use of ML in the Fight Against COVID-19

Mr Bogdan Laban, a master student at the University of Donja Gorica, defended his Master thesis titled “Practical Use of Machine Learning in the Fight Against COVID-19”. The thesis was done at the Master academic studies “Information Management in Health Care” at the Faculty of International Economics, Finance and Business, under the supervision of his mentor prof. dr Tomo Popovic.

Mr. Bogdan Laban defended his MSc thesis on the use of ML to fight COVID-19

The presentation included interested real-life examples

ABSTRACT – This study aims to find applicable combination of recent AI technologies and its uses during the COVID-19 pandemic. As AI has shown great improvements in the last few years, with many new feature-filled tools out on the market, it is almost certain that AI can help find ways to circumvent the dangers of COVID-19, in the form of noticing it, and preventing it. Authors cited in this work are AI professional, specializing in neural networks and their appliances in real world, in the form of visual processing and understanding, which will provide me with a solid hypothesis to base my master thesis on.

As this pandemic still has a solid grasp on the world, especially on less developed countries like Montenegro, I believe it makes a great testing ground for attributes like – percentage of people abiding pandemic specific laws (social distancing, mask wearing, etc.). Unfortunately, people of Montenegro are very divided on this issue, ranging from avid abiders, to those who completely ignore all safety procedures, even the most basic ones. As such, I believe this study will provide me with an insight into the right tools and tests for the residents of Montenegro, which vary a lot.

Another Master Thesis in Applied Machine Learning

20.03.2022.27.10.2022.

Evaluation of trends in jobs and skill‑sets using data analytics: a Case study, by Dr Armin Alibasic et al.

Member of the EuroCC Montenegro National Competence Center team, Dr Armin Alibasic together with co-authors Himanshu Upadhyay, Mecit Can Emre Simsekler, Thomas Kurfess, Wei Lee Woon and Mohammed Atif Omar conducted research published in the paper titled “Evaluation of trends in jobs and skill ‑ sets using data analytics: a Case study” in the Journal of Big Data. The article is available at the following link.

APSTRACT – A novel data-driven approach is developed to identify trending jobs through a case study in the oil and gas industry. The proposed approach leverages a range of data analytics tools, including Latent Semantic Indexing (LSI), Latent Dirichlet Allocation (LDA), Factor Analysis and Non-Negative Matrix Factorization (NMF), to study changes in the market. Further, our approach is capable of identifying disparities between skills that are covered by the educational system, and the skills that are required in the job market. Novel data-driven approach is developed to identify trending jobs through a case study in the oil and gas industry. The proposed approach leverages a range of data analytics tools, including Latent Semantic Indexing (LSI), Latent Dirichlet Allocation (LDA), Factor Analysis and Non-Negative Matrix Factorization (NMF), to study changes in the market. Further, our approach is capable of identifying disparities between skills that are covered by the educational system, and the skills that are required in the job market.

Dr. Alibasic published an article in the Journal of Big Data

18.03.2022.27.10.2022.

A Scientific Paper on Edge AI and Face Mask Detection

A paper titled “Face Mask Detection Based on Machine Learning and Edge Computing”, authored by I. Jovovic, D. Babic, S. Cakic, T. Popovic, S. Krco, and P. Knezevic, was presented at the 2022 21th International Symposium INFOTEH-JAHORINA. The paper was presented by a young researcher Mr Ivan Jovovic, Faculty for Information Systems and Technologies, on 18 March 2022. The paper discussed the use of machine learning for face mask detection and porting of prediction models onto the edge Ai platform. The effort was supported by the EuroCC Monteengro team link.

Presentation of the paper in the virtual online session of the conference

ABSTRACT – This paper describes research effort aimed at the use of machine learning, Internet of Things, and edge computing for a use case in health, mainly the prevention of the spread of infectious diseases. The main motivation for the research was the Covid-19 pandemic and the need to improve control of the prevention measures implementation. In the study, the experimentation was focused on the use of machine learning to create and utilize prediction models for face mask detection. The prediction model is then evaluated on the various platforms with a focus on the use on various edge devices equipped with a video camera sensor. Different platforms have been tested and evaluated such as standard laptop PC, Raspberry Pi3, and Jetson Nano AI edge platform. Finally, the paper discusses a possible approach to implement a solution that would utilize the face mask detection function and lays out the path for the future research steps.

The paper was presented at the IEEE 2022 21st INFOTEH-JAHORINA Conference

18.03.2022.27.10.2022.

A Conference Paper on Machine Learning in Agriculture

A paper titled “Detection of Plant Diseases Using Leaf Images and Machine Learning”, authored by A. Suljovic, S. Cakic, T. Popovic and S. Sandi, was presented at the 2022 21th International Symposium INFOTEH-JAHORINA. The paper was presented by a young researcher Ms Almira Suljovic, Faculty of Applied Sciences, on 18 March 2022 in the paper presentation session. The paper discussed the use of machine learning for detection of plant diseases that could be used in agriculture. The effort was supported by the EuroCC Monteengro team link.

Paper presented in the afternoon session by Ms Almira Suljovic

ABSTRACT – Prevention and early detection of plant diseases is one of the main issues and challenges in agriculture. Farmers spend a lot of time observing and detecting diseased plants, often by looking at and analyzing plant leaves. Inadequate handling of plant disease such as late detection or the use of wrong pesticides often causes damage to crops, which causes a deterioration in the quality of food. This problem could be addressed using artificial intelligence and machine learning to detect plant diseases by processing digital images of leaves. As the leaf is the best indicator of whether the plant is healthy or not, by applying machine learning we can create predication models to detect the condition of the leaf in a shorter period of time and possibly prevent or reduce the losses. This paper describes experimenting with Detectron2 software library and Faster R-CNN neural network in order to detect the condition of the leaf. A dataset containing 6407 images was used to train the model. The original dataset has been extended by augmenting images using the RoboFlow tool. The experimentation and implementation was done using Google Colab, environment designed for cloud computing and machine learning development.