Zexuan Wang

I am currently a graduate student at Georgia State University, majoring in Privacy and Security. I have taken several courses in machine learning and am very interested in AI. Previously, I worked in China for one year, where I developed and maintained a system, giving me a solid foundation in both front-end and back-end development.

Coding preprocessing of virtual reality video   02/2020~06/2020

• The project aims to improve the encoding efficiency of virtual reality video content.
• Preprocessed virtual reality videos using Python and OpenCV.
• Encoded videos using the HEVC encoder and calculated PSNR values for effective regions to validate the method's effectiveness.

An OKR management system   04/2021~05/2022

• The primary focus was to enhance the system to meet evolving customer demands.
• Redesigned the existing system with a front-end and back-end separation approach. Developed the back-end using Spring Boot, connected to the database with MyBatis, and built the front-end using Vue.
• Redesigned the front-end UI using Element components.
• Utilized ECharts components to present data.
• Redesigned all modules of the system and completed the system update without modifying the database structure.

Enhancing contrast for object classification   10/2023~12/2023

• The project's goal is to test various preprocessing methods to determine their effectiveness in improving the accuracy of image classification, a total of 1,500 images of goldfish and salamanders were used.
• Increased the image contrast of our dataset using various preprocessing methods, including gamma correction, image subtraction, histogram shift, and soft thresholding.
• Tested the image recognition accuracy using AlexNet.

An encryption method based on AES   10/2023~12/2023

• Designed a program that encrypts input characters using the AES algorithm, then replaces the results using a hash table. The program allows encoding and decoding to obtain the original and encrypted information.
• The purpose of this project is to conceal certain information to ensure the confidentiality of conversations.

Training a DNN model for image classification   02/2024~04/2024

• The purpose of the project is to train models on a cloud platform in a distributed manner.
• Utilized two deep neural network models, ResNet50 and VGG16, to test the CIFAR-10 dataset.
• Employed Azure as the cloud platform to set up virtual machines to facilitate the training environment, and TensorFlow to support distributed training.