AI Engineer Intern at FPT Software — traffic accident prediction research involving labeled visual data and YOLO model fine-tuning.
OpenCV AI Kit (OAK)
Video
Object DetectionLabeled and processed visual inputs to support supervised learning for an AI traffic-accident prediction pipeline. Prepared training data derived from dashcam video and used it to fine-tune detection models for risk-relevant features. Conducted evaluation of model performance and proposed improvements to increase prediction accuracy. • Labeled and processed images extracted from dashcam/video inputs. • Fine-tuned YOLO-based object detection to extract accident-risk features. • Evaluated model performance and analyzed results. • Proposed improvements to increase prediction accuracy.
2024 - 2024