Unsupervised Speech Recognition
I implemented unsupervised speech-to-text transformation with the wav2vec-U algorithm, enabling speech recognition for low-resource languages without sufficient training labels.
I implemented unsupervised speech-to-text transformation with the wav2vec-U algorithm, enabling speech recognition for low-resource languages without sufficient training labels.
I developed real-time automated object tracking applications for both template matching and motion detection, using the Lucas-Kanade algorithm with template drifting correction.
I conducted real-time image and video AR projections through interest point matching and homography estimation. I also made a panorama generation program that stitches photos together.
I reconstructed 3D objects from stereoscopic image pairs by computing the essential matrix, fundamental matrix, and triangulation using the eight-point algorithm.
Luis, Andy and I developed a simplified throwing game with different environments and modes. It can simulate the launch, study the motion, and animate the trajectory of projectiles.
We developed a Python program that allows users to search for a good, returns available options at nearby grocery stores, and compares their nutritional information through web-scraping.
I made a user-interactive web app with 3 investment modes, which can forecast cumulative returns and graph the growing trends given user input.
I created an R package that tests the validity of inputs, calculates the probability based on given parameters, graphs the distribution and displays a summary of skewness, kurtosis, etc.
I constructed shot charts, calculated the 2-pointer, 3- pointer and overall effective shooting rate, and derived the most suitable positions for five players, respectively.
I solved 56 SQL problems available on HackerRank and organized my solutions to each problem with their difficulty levels and tags.