What is it about?
In recent years, license plate prediction has advanced with machine learning, enabling accurate automated identification of vehicle plates for various applications, including law enforcement and surveillance. This study addresses the problem of effectively anticipating variations in traffic flow on metropolitan road networks, which are characterized by intermittent patterns and intense short-term volatility. DSOA-LSTM integrates Dove Swarm Optimization with Attention-based LSTM to improve traffic volume estimates using data analysis that takes into consideration spatial-temporal correlations. The attention mechanism in temporal correlation modeling demonstrates the importance of historical observations in the LSTM model. DSOA enhances license plate prediction by iteratively modifying parameters, emulating collaborative optimization, and is inspired by dove flocking behavior. The results showed that DSOA-LSTM is superior, with low Root Mean Squared Error (RMSE) of 6.55 and Mean Absolute Error of 4.65 and improved stability throughout several prediction phases, particularly in capturing intense short-term variations in traffic volume.
Featured Image
Why is it important?
This study introduces DSOA-LSTM, merging Dove Swarm Optimization with Attention-based LSTM, to enhance traffic volume prediction in metropolitan areas, crucial for law enforcement and surveillance. By leveraging spatial-temporal correlations and iterative parameter optimization inspired by dove flocking behavior, DSOA-LSTM demonstrates superior accuracy with low RMSE and improved stability in capturing intense short-term traffic variations.
Perspectives
Read the Original
This page is a summary of: License Plate Recognition using Attention-LSTM with Dove Swarm Optimization Algorithm, November 2023, Institute of Electrical & Electronics Engineers (IEEE),
DOI: 10.1109/iciics59993.2023.10421654.
You can read the full text:
Contributors
The following have contributed to this page