The concept of correlation pervades every scientific field. It is the measure of similarity or dependence between multivariate data, causal or not. In the following project, we introduce a type of correlation called cross-correlation and explore its applications. The project will be implemented in Python version 3.8, where the current repository contains all code, datasets, and documentation.

The project is divided into three main sections. The first section, introduces the mathematics of cross-correlation, its basic applications and a programmatic implementation. The second section is an investigation of stereo vision, which is the derivation of depth from multiple images. In this section, we will cover procedures including dot detection algorithms, calibration models, image comparison and 3D reconstruction. The final section explores an application in the field of finance, called a stock correlation network. In this section, we will implement algorithms to gather historical price data, calculate stock correlation matrices, and construct minimum spanning tree networks to represent stock correlation.

Main document: https://github.com/jkfids/cross-correlation/blob/main/docs/Report.pdf