Generative Yolov8-Based Deep Learning Framework for Real-Time Video Segmentation and Object Tracking in Multimedia Applications

View Full Paper

Ask AI

Bookmark

Authors

Discussion

This framework demonstrates improved object tracking and segmentation in video data, highlighting its performance and efficiency.

The model achieves 35 FPS real-time inference speeds and an mIoU of 82.3%.
It employs a generative adversarial network to enhance boundary precision and maintain object identities.
Evaluations on the DAVIS and MOT20 datasets show superior performance under challenging conditions.
The framework excels in handling occlusion, fast motion, and cluttered backgrounds.

Deshpande et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark