Description of Invention
This study aims to address the class imbalance problem through various approaches. Initially, a simple variation of oversampling technique is employed, wherein data from multiple classes were combined to increase the number of samples in underrepresented classes. To facilitate this process, the 18 vehicles types are divided into six groups, aggregating automobiles like the mini, sedan, SUV, and pickup under the "car" category. Apart from that, the use of attention mechanisms in deep learning is explored to overcome the class imbalance issue. By leveraging the attention mechanism, the model can focus on important features while disregarding less significant ones, thereby potentially mitigating the impact of class imbalance on model performance.