TY - GEN
T1 - “Unveiling the Invisible”
T2 - Deep Learning-based Semantic Segmentation for Analyzing Activity Patterns
AU - Kaur, Gurkiran
AU - Zhang, Li
PY - 2024/7/30
Y1 - 2024/7/30
N2 - The ubiquity of internet-enabled devices has led to a rapid increase in the use of connected cameras for real-time monitoring, creating a high demand for (automated) visual data analytics across various industries. The prospect of automating visual data analysis to drive positive change involves extracting actionable insights from data that will inform decision-making processes, improving efficiency, and contributing to evidence-based strategies across diverse applications and industries. This research explores and compares well-known semantic segmentation models such as DeepLabV3+ and UNet, determining the best-suited for use in a visual analytics and scene understanding, culminating in a proof of concept program capable of automating video analysis, plotting detections, average trajectories, and identifying outliers.
AB - The ubiquity of internet-enabled devices has led to a rapid increase in the use of connected cameras for real-time monitoring, creating a high demand for (automated) visual data analytics across various industries. The prospect of automating visual data analysis to drive positive change involves extracting actionable insights from data that will inform decision-making processes, improving efficiency, and contributing to evidence-based strategies across diverse applications and industries. This research explores and compares well-known semantic segmentation models such as DeepLabV3+ and UNet, determining the best-suited for use in a visual analytics and scene understanding, culminating in a proof of concept program capable of automating video analysis, plotting detections, average trajectories, and identifying outliers.
UR - https://www.worldscientific.com/doi/10.1142/9789811294631_0051
U2 - 10.1142/9789811294631_0051
DO - 10.1142/9789811294631_0051
M3 - Conference contribution
SN - 978-981-12-9462-4
T3 - World Scientific Proceedings Series on Computer Engineering and Information Science
SP - 411
EP - 418
BT - Intelligent Management of Data and Information in Decision Making
ER -