Projects
Face Detection System:
● Developed real-time face detection pipelines for video streams and image-based inputs.
● Integrated detection into FastAPI services for downstream analytics and recognition tasks.
● Optimized performance for GPU-based inference in Dockerized environments.
Face Recognition System:
● Designed and implemented an end-to-end face recognition system using ArcFace-based embeddings.
● Integrated detection, tracking, and identity matching for real-world surveillance and access scenarios.
● Focused on robustness, scalability, and real-time performance.
Vehicle Detection System:
● Built vehicle detection pipelines using YOLO-based models with GPU acceleration.
● Processed video streams and recorded footage with multithreaded frame handling.
● Prepared structured outputs for analytics and storage.
License Plate Recognition System:
● Implemented license plate detection and OCR-based recognition pipelines.
● Applied validation logic to ensure plates are detected before storing vehicle data.
● Sent structured inference results to a backend service via FastAPI; downstream components handled persistence (CSV and PostgreSQL) and further processing.
People Counting System:
● Developed people detection and tracking pipelines for estimating pedestrian volume and crowd density.
● Applied temporal aggregation to count people flow in urban environments.
● Used outputs to support traffic, footfall, and public space analysis use cases (e.g., streets in Jakarta).
Vehicle Counting System:
● Developed vehicle detection and tracking pipelines for estimating traffic volume on urban streets.
● Applied tracking, filtering, and temporal aggregation for vehicle counting and flow analysis.
● Improved inference stability for long-running video streams used in continuous traffic monitoring (Jakarta streets).
Cable-Label Detection System (Current Project):
● Currently developing a specialized object detection model for cable and label identification.
● Focused on industrial and inspection-oriented use cases.
● Working on dataset preparation, annotation strategy, and model optimization.
