HAO YIN

Mainly responsible for visual perception algorithms for automated parking systems - development and optimization of detection models

• Data cleaning, labeling, and diversity training data production;
• Model training and testing, code debugging;
• Establishing a perfect quantization, TDA4 board end self-test process, a script to complete all operations;
• Development and optimization of the 4bit car parking network on zu2 platform;
• For the project testing, analysis, recording, and model optimization;
• Data deep mining, tested model KPI for different parking scenarios, and forward optimized the model.