Year: 2020
Type: dissertation
Abstract: Machine Learning is computer program in which has the capability to learn from experience acquired from performing some task with some performance measure. Deep learning, a subfield of ML has progressed to the stage of being able to classify objects and recognize human activity [2], [3]. Human action recognition (HAR) in videos is a difficult and challenging problem [8] that has been widely studied in computer vision [9]. Convolutional Neural Networks (CNNs), developed based on the brain???s visual cortex, is the ideal architecture for HAR due to its high performance on complex visual tasks. \nMachine Learning research has predominately been focusing on learning algorithms in the past few years to invent ML models and applications where ML can be applied. In recent years, FPGAs have been used to deploy ML algorithms for training and inference. A PYNQ-Z1 board is a python-programmable FPGA-based System-on-Chip (SoC) that is used to deploy ML algorithms for video classification of HAR.
Author: Jonathan Hoong
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3205814928