Transformation models for images

• Main Author: Cheah, Ann Chee
• Other Authors: Chua Chin Seng
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/77318
• Institution: Nanyang Technological University

With relevance to Image Processing or Computer Vision Software Development, Transformation Models of Images is a technique that involves combining multi-images by discovering the relationship in overlapping areas of the images to create a detailed output image of a greater field of view and/or a higher image resolution. Through this project, the study provides a deeper understanding of the theory behind the execution of codes to create a stitched image from simple to complex models with the aim of making it applicable for practical applications. There are many applications that use the transformation model technique such as medical and geographical imaging and not limited to applications for aesthetical purposes. The programming language used in the project is Python on IDE PyCharm. Using OpenCV library and two main algorithms, namely SIFT, a feature detector that extracts unique point of interest from an algorithm to find the descriptor of each of such key point and RANSAC, an iterative method to detect any outliers by estimating the parameters of the mathematical model from a set of observed data. The difficulties encountered throughout this project was the ability to truly understand, investigate and to keep up with the time phase given. Troubleshooting with any inconsistent results, errors or limitation is one of the main challenges which is also a continuous experimental process in improvising the codes and obtain the goal of the project. Nevertheless, this project provides a great learning process and a better understanding of Image Transformation by gaining insights into the background procedures and the underlying theories.