A systematic evaluation of deep learning models for vehicle classification and counting using numerical data
This study provides a comprehensive baseline evaluation of five deep learning models 1D Convolutional Neural Network (1DCNN), Long Short-Term Memory (LSTM), 2D Convolutional Neural Network (2DCNN), Recurrent Neural Network (RNN), and Autoencoder on two distinct dataset types: a synthetic dataset and...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Proceeding |
| Language: | en |
| Published: |
IEEE
2025
|
| Subjects: | |
| Online Access: | http://ir.unimas.my/id/eprint/51437/3/A%20Systematic%20Evaluation.pdf http://ir.unimas.my/id/eprint/51437/ https://ieeexplore.ieee.org/abstract/document/11330380 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This study provides a comprehensive baseline evaluation of five deep learning models 1D Convolutional Neural Network (1DCNN), Long Short-Term Memory (LSTM), 2D Convolutional Neural Network (2DCNN), Recurrent Neural Network (RNN), and Autoencoder on two distinct dataset types: a synthetic dataset and a hybrid dataset. Structured numeric data exists in two separate datasets: The synthetic dataset spans from 10K to 100K points and the hybrid integration of synthetic data with real-world data also falls within this parameter range. Results from experiments show 1DCNN excels at numeric data processing due to its ability to deliver superior results with increased efficiency rates beyond competing models. The experimental results of this study validate the compatibility and efficient data handling abilities of 1DCNN across various large numeric
datasets, resulting in reduced computational complexity. This
study provides essential knowledge about deep learning model
capabilities in numeric data processing so future AI systems and data-driven decisions can develop further. |
|---|