EUGENE HSUE

Dynamic Cell Culture Manufacturing Associate at Amgen with expertise in aseptic techniques and production operations. Proven ability to enhance process efficiency and ensure compliance with GMPs. Strong collaborator and problem solver, adept at monitoring critical parameters to optimize cell line viability and drive continuous improvement initiatives.

• Eigenface-Based Facial Recognition System, https://github.com/Str0nkyK0ng/UCLA-S23-Math155, Tools & Technologies: MATLAB R2023a, Singular Value Decomposition (SVD), Eigenfaces, Principal Component Analysis (PCA), Image Processing, Algorithm Implementation: Utilized SVD for extracting facial features, forming a robust Eigenface-based recognition process., Data Handling & Pre-processing: Managed image datasets, normalized images, and computed key eigenvectors for facial feature distinction., Performance Analysis: Tested system against different image transformations (blurring, rotation, intensity adjustments) to simulate real-world scenarios., Innovation & Problem-Solving: Addressed challenges in image processing and algorithm optimization, enhancing system's adaptability to environmental variations.
• AI-Powered Gait Analysis for Neurodegenerative Diseases, https://github.com/norfork37/NTU_inpaintandpose, System Development: Implemented an end-to-end AI interface integrating Detectron2, RunwayML Stable Diffusion, and Gradio with Hugging Face for seamless user interaction., Data Preparation & Analysis: Prepared and processed a dataset of 2049 videos, focusing on patients with varying stages of Parkinson's disease., Advanced Segmentation & Tracking: Utilized Detectron2 for precise segmentation and tracking of patients, effectively handling video frames with multiple individuals., Inpainting Techniques: Applied multiple inpainting methods (e.g., flow-guided video inpainting, OpenCV, Stable Diffusion) for background individual removal, enhancing the clarity of patient movements in videos., Interface Building: Built a user-friendly interface using Gradio and Hugging Face, integrating AI models for accessible and efficient analysis.
• Graph Analysis in Epidemiology, https://github.com/norfork37/Graphs-cse100, Graph Class API Development: Engineered a Graph class in C++ to represent undirected graphs, encapsulating essential graph operations and properties., Graph Construction & Analysis: Enabled construction of graphs from viral sequence data, calculating pairwise distances under the TN93 DNA evolution model and identifying transmission clusters., Algorithm Implementation: Implemented critical graph algorithms including unweighted and weighted shortest path (BFS and Dijkstra's algorithm), connected components detection with thresholds, aiding in understanding virus spread dynamics., Efficient calculation of the smallest connecting threshold between individuals using a Disjoint Set data structure., Data Handling: Developed functionality to parse and handle edge list data in CSV format, representing complex epidemiological data sets.
• Huffman Coding for File Compression and Uncompression, https://github.com/norfork37/HuffmanCoding, Huffman Code Construction: Devised a system to construct Huffman codes from the contents of the input file, ensuring optimal compression based on data frequency., Compression Algorithm: Created the compress program to encode the original file using the constructed Huffman codes. Output a compressed version of the file with a header to facilitate uncompression., Uncompression Algorithm: Developed the uncompress program to reconstruct the Huffman code by reading the header of the compressed file and decode the compressed file and restore it to its original form., Data Integrity: Ensured that the uncompress program accurately restores the original file, verified by running diff between the original and uncompressed files with no differences reported.