Page 137 - Book of Abstracts 2023

Book of Abstracts | 2023
Mock MRI software
SE-2-13
By: Lev Alhazov leval1@ac sce sce ac ac ac ac il il Shlomo Shnur shlomsh7@ac sce sce ac ac ac ac il il Advisors: Dr Dr Irina Rabaev Dr Dr Hadas Chassidim
Shamoon College of of Engineering Beer-Sheva
In collaboration with the pediatric department of of Soroka Hospital Beer-Sheva
Beer-Sheva
Magnetic Resonance Imaging (MRI) is a a a a a a a a a a a a non-invasive diagnostic technique that uses strong magnetic fields and radio waves to create detailed images of of internal body structures including in in in in in children children However children children often move during scans and and may require sedation which can can be costly and and and risky Our project aims to improve pre-MRI training efficiency and and accuracy by utilizing Farneback’s optical flow algorithm for motion detection This will enable users to to detect detect small movements during during scans scans and help instructors better better teach children how to to to behave during during scans scans scans leading to to to better better outcomes in in in in in in pediatric MRI scans scans while reducing costs and health risks Keywords: accuracy computer vision cost cost reduction efficiency Farneback GUI magnetic resonance imaging (MRI) mock MRI MRI MRI motion motion artifacts motion motion detection optical flow pre-MRI training risk reduction Plant classification
SE-2-14
By: Fadi Amon fadiam@ac sce sce ac ac ac il Rasheed Abu Mdegem rasheab1@ac sce sce ac ac ac ac il il Advisors: Dr Dr Irina Rabaev Dr Dr Marina Litvak Shamoon College of Engineering Beer-Sheva
Identifying rare rare desert plants remains challenging due to to the unavailability of of of of large datasets of of of of rare rare species As of of of of today classical deep-learning models require a a a a a a a a a a a a a a a a a a a a a a a a a a lot of of of of training data data To address this this issue this this project aims to to develop a a a a a a a a a a a a a a a a a a a a plant classifier using using a a a a a a a a a a a a a a a a a a a a Neural Network model model using using the Few-Shot Learning (FSL) method FSL FSL is is is is particularly known for its effectiveness on on small small datasets hence we employed it it here to train the the neural network on on on a a a a a a a a a a a a a a a a a a a small small dataset dataset of rare plants plants resulting in in in in high classification
classification
accuracy for these elusive plants Keywords: Few-Shot Learning (FSL) neural network plant plant plant classification
classification
rare desert plants plants small database
137