I am a thild-year undergraduate student in the Software Engineer at the College of Computer and Information Science, Southwest University. I’m open to any kinds of collaboration. Please feel free to contact me directly through email.
📧 [email protected] | 📱 (+86)151-8699-9278 | 🐱 GitHub
Guiyang No.1 High School 
| Sep 2019 - Jun 2022
Southwest University (SWU)
Bachelor of Engineering in Software Engineering | Sep 2022 - Jun 2026
GPA: 83.086/100 | IELTS: 6.0
Comprehensive Analysis of Fruit Detection, Ripeness Assessment, and Mass Estimation Using YOLOv7, ResNet, and ViT Models
Zhihan Guo, Minqi Liu
IEEE 3rd International Conference on Electrical Engineering, Big Data and Algorithms 2024
[Webpage]
- Third-class Scholarship | Southwest University (Top 17.5%) | Oct 2024
- Silver Award | 26th China Robotics and AI Competition (New Quality Productivity Innovation Design) | May 2024
- Bronze Award | 26th China Robotics and AI Competition (AI Innovation) | Apr 2024
- Bronze Award | 15th Langiao Cup Digital Tech Innovation Competition | Apr 2024
- Silver Award | 14th China National Undergraduate "Innovation, Creativity and Entrepreneurship" Challenge | Apr 2024
- Honorable Mention | Mathematical Contest in Modeling | Feb 2024
- Meritorious Award | 12th Certificate Authority Cup Int’l Mathematical Contest in Modeling | Nov 2023
- Silver Award | Gauss Cup National Mathematical Modeling Challenge | Nov 2023
- Silver Award | 18th National University Students Intelligence Car Race | Nov 2023
- Target detection and counting: YOLOv7 model was used to achieve apple target detection, 200 samples were detected using SPSS, and the number of apples in a single image was counted by the number of detection frames, and a histogram of the distribution of the number of apples generated was obtained.
- Maturity grading assessment: Definition of three stages of maturity criteria, calculation of maturity based on red/yellow/green pigment weights, and development of a HSV colour space analysis model.
Map detection boxes to apples Histogram of apple count distribution by using Histogram of maturity distribution of apples
- Mass estimation algorithm:Introduce the coin reference to establish the pixel-mass conversion model ,derive the mass calculation formula, then estimate the mass of a single fruit by the red pixel area.
Apple ripening process chart Histogram of maturity distribution of apples
- Comparative Analysis of Classification Models: Constructing ResNet and ViT Dual Model Classification Frameworks, Validation Set Accuracy Comparison Experiments.
Comparison of resnet model and vif model results
Jan 2024 - Feb 2025
- Integrated diffusion models and text-driven tech to develop a military chess sandbox generation system.
- Built a high-quality dataset through data cleaning, annotation, and augmentation (rotation/scaling).
- Optimized model performance via hyperparameter tuning, network structure refinement, and regularization.
- Collaborated with military experts to validate results; Authored 10,550-word proposal for national project approval.
Nov 2023 - Mar 2024
- Designed wheel-track composite structure for terrain adaptability (slopes, stairs, elevators).
- Implemented GPS/BeiDou navigation and RVIZ for real-time 3D motion tracking.
Flowchart of the intelligent elevator system simulation model
- Integrated TOF radar and Molex laser tech for obstacle avoidance; Created CAD 3D diagrams and demo videos.
- Wrote 78-page technical manual; Secured provincial-level project approval.
rendering model
Feb 2024
- Captured scoring patterns in tennis matches and analyzed key factors influencing results using randomness analysis.
- Preprocessed match data via missing value imputation, outlier detection, and feature standardization; managed datasets with MySQL.
Visualization of datas(players' Momentum Score) Crossplot of GAME win/loss scenarios and momentum score
- Built a momentum scoring model with TOPSIS algorithm, normalizing 8 attributes (e.g., serving speed, hitting accuracy).
- Conducted statistical analysis with SPSS, applied The Run test to prove correlation between momentum levels and match outcomes.
Result of the run test
- Developed a Markov model to predict momentum states, derived momentum probability expectation formulas.
- Validated model accuracy through Markov steady-state distribution tests and proposed data-driven strategies for match optimization.
Markov-Based Tennis Momentum State Prediction Model running results Conceptual Diagram for Establishing the Model
Nov 2023
- Designed a standardized evaluation framework for telescope performance in low-light environments.
- Created a resolution model integrating telescope parameters, optical characteristics, and human eye perception, calculating correction factors for dim conditions.
- Innovated a twilight factor algorithm based on signal-to-noise ratio, exit pupil intensity, and resolution patterns.
- Developed distinct algorithms for binoculars (human-eye observation) and CMOS video lenses, enabling standardized device comparisons.
- Provided weighted nonlinear processing to compute twilight factors, addressing limitations of traditional models.
- Delivered actionable recommendations for telescope selection in astronomical observation scenarios.
Image of the signal-to-noise ratio equation Eye brightness sensory map The application of the new dusk coefficient in 7 lens
Guizhou Zhicheng Technology Co., Ltd.
Big Data Analytics Engineer | Jul 2024
- Analyzed bus station data using Hadoop/Spark to identify passenger boarding/alighting patterns.
- Developed transfer coefficient algorithm for route transfer behavior analysis.
- Built time-series models for travel volume forecasting; Visualized data with Tableau/Matlab.
Hong Kong University | Jun 2025
- Learned how to use generative AI to generate images and videos.Used generative AI in education to assist people in learning
Chinese idioms.
- Learned about the Python programs on Colab and how to use prompts to enhance the quality of AI-generated images and videos.
- Analyzed the impact of different prompts on the AI and compared the differences between different AI platforms.
- Accomplished all the workshop and delivered a 15-minute presentation in English.
The Chinese University of Hong Kong | Feb 2024
- Researched SenseTime’s AI risk assessment and Bowtie Insurance’s digital marketing strategies.
- Proposed blockchain integration for food supply chain traceability (Qian Da Ma Supermarket case study).
- Delivered 34-minute English presentation (Grade: A-).
Xingke Debate Team, SWU | Member | Sep 2022 - Jul 2024
- Organized and hosted university-level debate competitions.
- Managed event planning and cross-team communication.
