Deep Neural Network-based Approach for Accurate Vehicle Counting

Abstract

In highway management, intelligent vehicle detection and counting are becoming increasingly important as an accurate estimation of traffic density on road congestion reduction. Traffic density estimation is affected by the difficulties of perspective distortion, size change, significant occlusion, and background interference in traffic images. To address the previous issues, this article develops a novel model that enhances the quality of estimating traffic density. The efficientNet fine-tuning architecture is used then, followed by the development of seven dilated convolutional layers to extract the deeper features in the images that maintain the output's resolution to generate a high-quality density map. Finally, the vehicle count will be calculated from the high-quality density map. The experimental results indicate that the suggested approach significantly enhances the accuracy of traffic density estimation compared to the existing ones. It achieves 5.23 as a mean absolute error (MAE) on the TRANCOS dataset.

Performance Analysis of End-to-End Neural Coreference Resolution in English and Indonesian Texts

“…Mention Score [20] (s m ) is the value of how big a span is a mention. This value is calculated using an Artificial Neural Network [27], [28]. Only the few spans that have the highest value will be taken as a mention, the rest will be prune.…”