Projects
Here are some brief details of a few machine learning and/or computer vision projects outside of my research in fingerprint and face biometrics.
Predicting National Happiness
Concepts: Machine Learning
Ridge regression analysis for predicting national happiness from data presented in the World Happiness Report and CIA Factbook. Incorporated data augmentation and preprocessing techniques such as Matrix Imputation, Z-score normalization, clustering analysis, etc.
Predicting Online News Popularity
Concepts: ML and Pattern Recognition
Evaluated 8 different ML classifiers (RBF-SVM, MLP, Random Forest Decision Tree, AdaBoost Ensemble of Weak Learners, Bayes classifier w/ non-parametric (KNN and Parzen Window Estimators) and parametric (Gaussian) class-conditional distributions in combination with various preprocessing and feature selection techniques.
Meal Recognition
Concepts: Deep Learning and Computer Vision
Deep learning applied to meal recognition involving transfer learning to overcome a limited training dataset (~1000 images w/ 14 classes). Additionally, performed object detection via pseudo labeling of bounding-boxes.
