Joy Hu

TOF Depth Image Processing and System Optimization

• Improved color mapping and distance calculation methods to enhance color accuracy and depth precision in high-density point cloud data.
• Utilized shared memory and Linux IPC, leveraging RingBuffer to optimize data flow efficiency, resulting in a 550% improvement in processing speed.
• Processed 2D/3D point cloud data using PCL and rendered visualizations with three.js.
• Co-inventor of a related patent (Patent No. TW113200467).

Production Process Optimization Contributions

• Developed a Python-based cargo stacking simulation system that automatically selects the optimal stacking method, enhancing packaging efficiency and integrating with the MES system to boost production line efficiency.
• Developed a C# label generation system to streamline processes, reduce errors, and integrate with the MES system to ensure label synchronization with production processes.

Bird Identification and Statistics System Development

• Developed a bird identification system using Flask for real-time identification and quantity statistics.
• Employed YOLO V9 and EfficientNet V2 for object detection and classification, supporting RTSP video capture and PTZ functionality for IPcam streaming.
• Enabled real-time data display and the ability to query by time range or species.

Facial Recognition Plugin Development

• Integrated Lilin facial recognition plugin with Innovatrics SmartFace Stream Processor.
• Processed facial recognition data via WebSocket and REST API, integrating IPcam streaming for real-time data capture and processing
• Implemented alert systems for restricted areas, watchlists, and prolonged stay notifications, with email notifications for users.
• Authored technical documentation to assist with system maintenance and optimization.

Technical Optimization and Process Automation

• Performed memory analysis using Valgrind, resolved memory leakage issues, enhanced system stability, and successfully ported the optimized code to embedded platforms through Makefile and cross-compilation.
• Implemented Docker and Jenkins for automated building and packaging, reducing deployment cycles and improving development and operational efficiency.

PTZ and Spotlight Integration System Development

• Led the development of a PTZ camera and spotlight integration system, using TCP/IP and WebSocket for real-time data synchronization.
• The project was showcased at the 2025 ISC WEST International Security Technology Exhibition in the U.S.

4th Generation Ticketing System of Taiwan Railways Administration

• Designed microservices based on Spring Boot, implementing fare calculation, ticket reissue services, and gate validation systems to enhance system scalability and maintainability
• Integrated Spring Data JPA and Oracle to implement dynamic fare calculation and ticket reissue logic, supporting real-time adjustments and complex ticketing rules
• Utilized JUnit and Mockito for testing, combined with Jenkins and SVN for automated building and deployment.