LabelHelper
03/2024 - Present | GitHub
Master Thesis: This research investigates the use of an OpenAI API chatbot to educate senior citizens on utilizing Roboflow for image labeling. The study aims to assess the chatbot's effectiveness in teaching, efficiency in imparting knowledge, and the accuracy of image labeling performed by seniors post-training. This approach seeks to bridge the digital divide by making advanced AI tools accessible and user-friendly for the elderly population.
Pseudo-LiDAR by StereoVision
09/2023 - 01/2024 | GitHub |
The Pseudo-LiDAR project leverages stereo vision techniques with two simple cameras to recognize objects and calculate depth. Using Roboflow for data preprocessing and SIFT (Scale-Invariant Feature Transform) matching algorithms, the project processes stereo images to identify objects and estimate their distances. This approach offers a cost-effective alternative to traditional LiDAR systems, providing accurate depth perception and object recognition capabilities for various applications.
Optimizing Pharmaceutical Portfolios Using Advanced FinTech Models
12/2020 - 05/2021 | PDF |
This project uses methods like the Equal Weighted Model, Monte Carlo simulation, and Sharpe ratio maximization to evaluate five key assets. By applying the Monte Carlo method, we developed optimized portfolios with minimum variance and maximum Sharpe ratio, revealing NVS as the largest component. These insights are valuable for investors in the pharmaceutical sector.