Human Activity Recognition (HAR) refers to automatic recognition of engagement of a subject in certain activities, so it effectively answers the questions of what is a subject doing and when, as well as assessing how well a subject is doing these activities. The foundation for HAR systems is activities data captured using sensors of various types, including body-worn inertial measurement units such as accelerometers and gyroscopes or other audio or visual sensors. In applications of HAR, such as healthcare, it is often important to capture the activities being performed by the subject, through a detailed recording of the daily activities, to achieve better understanding of the activity being undertaken by a patient for diagnosis of pathologies.
By leveraging advances in deep learning, researchers worked on developing novel approaches to resolve challenging feature selection and pattern recognition problems in various domains of signal processing, computer vision, speech recognition, natural language processing, and more.
This talk discusses application of different deep learning models, such as convolutional neural network (CNN), recurrent neural network (RNN), and others, to automate feature learning from the raw inputs in a systematic way for human activity recognition.
Dr. Nashwa El-Bendary is an associate professor at College of Computing and Information Technology, Arab Academy for Science, Technology, and Maritime Transport (AASTMT), Egypt. She is also the director of China-Arab States Technology Transfer Center at AASTMT (CASTTC-AASTMT).
She received her Ph.D. degree in Information Technology from Faculty of Computers and Information, Cairo University, Egypt, in 2008.
Her research interests include Internet of Things (IoT), Wireless Sensor Networks (WSN), Image processing, Signal Processing, and Deep Learning.
She has more than 70 scientific research publications published in prestigious journals, peer-reviewed international conference along with book chapters published in books of reputed publishers.
Dr. El-Bendary is the recipient of the 2014 UNESCO - ALECSO Award for creativity and technical innovation for young researchers in the Arab World and the 2015 The L’Oréal-UNESCO "For Women in Science" Levant & Egypt fellowship. Dr. El-Bendary is an editorial board member in Applied Soft Computing Journal, Elsevier. She co-organized and co-chaired a number of special sessions and workshops under the framework of distinguished international conferences as well as being invited as a speaker in several workshops and international conferences.