Abstract:
Gradient-based surface reconstruction has emerged as a research field in
computer vision, finding applications across various domains. This thesis aims to delve into
the current state-of-the-art techniques utilized in gradient-based surface reconstruction. It
has been observed that subtle variations in surface gradient can produce significant effects
on the output of these reconstruction algorithms. Therefore, this study explores the
influence of different surface features and parameters on various surface reconstruction
techniques.
Nonetheless, comparative studies to comprehensively understand the performance of
different gradient-based surface reconstruction methods under similar scenarios are
limited. To bridge this gap, this thesis analyzes the state-of-the-art techniques used in
gradient-based surface reconstruction. We selected five representative methods and
designed experiments that simulate diverse surface conditions, with the goal of measuring
their robustness and efficacy.
This review thesis will aid researchers in selecting the most suitable gradient-based surface
reconstruction method for their specific applications, since we present comprehensive
analyses highlighting the impact of various surface conditions and parameters on the
output of the selected reconstruction techniques.
