Abstract:
The development in the research field in the past few years have been seen in
Machine Learning (ML), deep learning and Artificial Neural Network (ANN). The
application of deep learning, especially Convolutional Neural Networks (CNN) is used
in image classification, image semantic segmentation, object detection in images, etc.
The CNN have already been used in automatic image classification systems. This
research focuses on images classification using CNN’s. Different design layers,
activation function, normalization, pooling, feature map optimization, and fast
computation of model are all included in Convolutional Neural Networks (CNN), which
provides better accuracy. The Convolutional Neural Network learns to represent the
images and a trains classifier is used to label images. In this work Convolutional Neural
Network is used for multi-class classification problems and is trained on the large
sample of the images which improves the butter accuracy of the model. The goal of this
thesis was to evaluate the results and obtain better accuracy of the model which is uses
convolutional neural networks in image classification. Dataset of the Landscape
Recommendation system has been acquired from the Kaggle repository. After preprocessing
on the dataset, different techniques named as, Data, augmentation, CNN, KFold
validation, and Modified VGG-16. K-fold cross-validation are applied by splitting
our dataset into training and testing sets. After model training, it is observed the
accuracy has been improved with the help of the proposed methodology using CNN.
Proposed technique achieved accuracy of 92% which is better than state of the art work.
