Improving Dog Breed Classification Performance through Stacked Deep Learning Models with Imbalanced Data

Image classification is a computer vision task that helps to classify the images based on various techniques with respect to feature extraction process. This research work is put together into two tasks. First, it aims to address the class imbalance issue inherent in the dataset through the utilization of data augmentation techniques. Secondly, the focus shifts towards employing deep learning (DL) methodologies for dog breed classification using stacked model. Stacked model is to combine the one or more DL techniques that helps to achieve good results. Evaluations are carried out using the following metrics such as accuracy, f1-score, precision, and recall. These metrics helps to analyse the performance of proposed stacked model in dog breed classification. Experimental result proves that the combined stack model with data augumentation techniques helps to achieves 88.7% result in the dog breed corpus.