Implementing some algorithms for image registration

Project of EE6553 (2020/12/22)

Abstract—These days, object registration and tracking in a sequence of images play a significant role in various areas. One of the methods in image registration is featurebased algorithms that in two steps were accomplished. The first step includes finding features of scene and object images. In this step, for reducing the sensitivity of detected features to the scale changes, scale-space is used. Afterward, we attribute feature points obtained in the first step, a description using brightness value around the feature points. In this project, five algorithms such as Binary Robust Invariant Scalable Keypoints (BRISK) and Scale Invariant Feature Transform (SIFT) are implemented in python. These algorithms could use in object detection and tracking.

Image registration, SIFT, BRISK, BRIEF, FREAK, ORB.

Document
GitHub

Programming Project of Machine Learning and Data Mining

Project of CS6735 (2021/04/20)

Conduct an experimental study on the following machine learning algorithms:
(1) ID3; (2) Adaboost on ID3; (3) Random Forest; (4) Naive Bayes; (5) K-nearest neighbors (kNN).

Document
GitHub

Time-Series Classification

Project of EE6563 (2021/04/20)

In this project, we present a review of the time series approaches for classification tasks including conventional machine learning algorithms and Deep Neural Networks. These approaches were implemented in verification mode.

Document
GitHub