Resource-efficient fine-tuning of large vision-language models for multimodal perception in autonomous excavators

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Analysis shows improved object detection and F1 score in autonomous excavators, suggesting fine-tuning might enhance safety monitoring.

Significant improvements in object detection and F1 score are reported in the analysis of autonomous excavators.
Fine-tuning with heuristics achieved an impressive F1 score of 78.83% on a challenging domain-specific dataset.
Evaluated using a unique dataset with 1,000 annotated frames and testing on 2,000 images from actual excavator scenarios.
These advancements indicate the potential for deploying advanced AI agents in construction settings for enhanced operational safety.

Nguyen et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark