Video Analysis for Event Detection in Smart Buildings Using Machine Deep Learning Algorithms

Abstract

Computer vision focuses on enabling machines to interpret and understand digital images and videos. It has numerous applications in various domains, including: Autonomous vehicles, Medical imaging, Surveillance and security, Robotics, Augmented reality, and many more. Event detection is a computer vision task that involves identifying and categorizing events or activities that occur in images. There are many potential applications, from surveillance systems that detect suspicious activities to monitoring systems that automatically tag and categorize images based on the events or activities depicted. Event detection in videos can be done using various techniques and algorithms from computer vision and machine learning. Some popular methods include: (i) Optical flow: Analyzing the motion of pixels between consecutive frames in a video to detect changes and events. (ii) Object detection: Detecting and tracking objects in a video by identifying their characteristics and movements. (iii) Feature extraction: Extracting and analyzing features such as color, texture, and shape from frames in a video to identify events. (iv) Deep learning: Applying convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to learn and detect complex patterns and events in videos. (v) Semantic segmentation: Segmenting different objects and regions in a video and analyzing their interactions and movements to identify events. (vi) Activity recognition: Recognizing and classifying different activities and actions in a video, such as walking, running, or jumping. These methods can be used individually or in combination with each other to improve the accuracy of event detection in videos. The objective of this thesis is to detect events in smart buildings from video obtained through surveillance camera. This thesis work focuses on analyzing both online and offline videos. newline Online video Analysis (Real Time Analysis): For Immediate detection of events in frames as they are captured. Offline video Analysis (Pre-recorded Video...