Here, I will share with you my learning and practical journey in the field of computer science, covering multi-dimensional experiences such as web development and computer vision algorithms.
I hope that through this platform, I can exchange technical insights with like-minded peers and grow together. Now, follow my perspective and embark on this journey filled with code and innovation!
Tongyun Huishi(铜韵慧识) — Ancient Bronze Intelligent Identification and Learning Platform
Empowering ancient bronze identification and learning through AI technology. It includes four modules: Ancient Mirror Appraisal, Bronze Encyclopedia, In-depth Bronze Study, and 3D Reconstruction. Additionally, a community platform has been built, allowing users to post, like, and collect content, with a large model function provided in the bottom right corner. (A national-level innovation and entrepreneurship project, served as the project leader, and has been used in teaching at the High School Affiliated to Renmin University of China.)
Multi-source Meteorological Data Visualization and Analysis Platform
Build a multi-source meteorological(historical data + real-time data) data visualization platform. It includes four functional modules: real-time weather query, historical meteorological data query, life index, and disaster warning. The front-end uses Vue3, the back-end uses SpringBoot, MySQL for data persistence, and ECharts for chart rendering.
Research on Ancient Bronze Classification and Dating Algorithms
By integrating features such as contour, vessel shape, and decorative patterns, an end-to-end multi-task model for the classification and dating of ancient bronzes was constructed, and a bronze artifact category-function matrix was designed to assist the classification task. Ablation experiments demonstrated that the classification accuracy reached 97.59% , and the dating accuracy reached 87.57% .
3D Reconstruction and Digitization of Ancient Artifacts
This project utilizes FFmpeg, Colmap, and the open-source Instant-NGP model to perform 3D reconstruction and digital processing of bronze artifacts exhibited at Shanghai University, generating high-precision 3D models.
Research on Panoramic Image Quality Assessment Algorithms
A subjective experiment was designed, organizing 16 participants to rate 3,200 viewport images, thereby constructing a dataset. Based on this, a weighted multi-viewport CNN algorithm was developed, with results showing that the algorithm achieved an SROCC metric of 0.9665, indicating excellent performance.
Email: 2295894399@qq.com