An Integrated Approach towards Efficient Image Classification Using Deep CNN with Transfer Learning and PCA

Abstract

In image processing, developing efficient, automated, and accurate techniques to classify images with varying intensity level, resolution, aspect ratio, orientation, contrast, sharpness, etc. is a challenging task. This study presents an integrated approach for image classification by employing transfer learning for feature selection and using principal component analysis (PCA) for feature reduction. The PCA algorithm is employed for reducing the dimensionality of the features extracted by the VGG16 model to obtain a handful of features for speeding up image reorganization. For multilayer perceptron classifiers, support vector machine (SVM) and random forest (RF) algorithms are used. The performance of the proposed approach is compared with other classifiers. The experimental results establish the supremacy of the VGG16-PCA-Multilayer perceptron model integrated approach and achieve a reorganization accuracy of 91.145%, 95.0%, 92.33%, and 98.59% on Fashion-MNIST dataset, ORL dataset of faces, corn leaf disease dataset, and rice leaf disease datasets, respectively.

Framework for deep learning based model for human activity recognition (HAR) using adapted PSRA6 dataset

“…The datasets which are used in HAR can be sensor- algorithms can be employed as classifiers [25]. The best error rate was achieved by faster training through non-saturating neurons for a large, deep CNN with five convolutional layers, followed by max-pooling layers and two globally connected layers [26].…”

Non-Facial Video Spatiotemporal Forensic Analysis Using Deep Learning Techniques

“…Instead, CNNs that are simple to create have become popular. Sharma and Singh [19] proposed a combined technique of image classification that employs transfer learning for feature selection and principal component analysis (PCA) for feature reduction. Capsule networks have now been created with impressive early results due to the introduction of the expectation-maximizing routing algorithm along with the dynamic routing algorithm [20].…”

Pollination Based Optimization: A New Global Optimization Algorithm and Its Application to Paddy Disease Detection

“…As referred to CNNs, hyper-parameters are the variables those control as to how the network is trained and how its structure is set up. CNN architecture makes use of a large number of hyper-parameters [19][20]. Domain/Technical knowledge is necessary to select the optimal hyperparameters manually.…”