Convolutional neural networks (CNN) are a boon to image classification algorithms as it can learn highly abstract features and work with less parameter. Overfitting, exploding gradient, and class imbalance are the major challenges while training the model using CNN. These issues can diminish the performance of the model. Proper understanding and use of corrective measures can substantially prevent the model from these issues and can increase the efficiency of the model. In this paper the conceptual understanding of the basic CNN model along with its key layers is provided. The paper summarizes the results of training the deep learning model using CNN on publicly available datasets of cats and dogs. Finally the paper discusses various methods such as data augmentation, regularization, dropout, etc. to prevent the CNN model from overfitting problem. The paper will also help beginners to have a broad comprehension of CNN and motivate them to venture in this field.
CITATION STYLE
Joshi, S., Verma, D. K., Saxena, G., & Paraye, A. (2019). Issues in Training a Convolutional Neural Network Model for Image Classification. In Communications in Computer and Information Science (Vol. 1046, pp. 282–293). Springer Verlag. https://doi.org/10.1007/978-981-13-9942-8_27
Mendeley helps you to discover research relevant for your work.