Deep learning approaches have been found to be suitable for the agricultural field with successful applications to vegetable infection through plant disease. In this chapter, the authors discuss some widely used deep learning architecture and their practical applications. Nowadays, in many typical applications of machine vision, there is a tendency to replace classical techniques with deep learning algorithms. The benefits are valuable; on one hand, it avoids the need of specialized handcrafted features extractors, and on the other hand, results are not damaged. Moreover, they typically get improved.
TopIntroduction
Deep Learning is a representation of learning with several features at different layers. These features are self-evolutionary and are produced after each iteration. The features re observed at each layer and are inter dependent on previous layer. Each feature may be an input for the next layer. It is an emerging technique and a procedure that incorporates ANN. Deep learning is one of the machine learning techniques that enables computers with lots of data and intermediate results that machine can analyze and provide justification for any answer.
Deep learning algorithm are used to apply on large amount of unsupervised data and provide different views of relationship or view of complex representation. Deep learning algorithms use the basic theory and concept of Artificial Intelligence and therefore these algorithms are able to emulate the human brain’s ability to analyze, learn and able to make decision even in the case of complex problems.
It has a wide variety of applications in different field of learning and experiment. A model empaneled with deep learning methodology can perform classification tasks over images, texts, video or a combination of these. The models are trained by large amount of unsupervised data, follow the NN architecture and hence deep learning models are able to achieve accuracy at significant level that consider different aspects of result and accepted by larger end users.
Since it works on huge amount of data, a substantial computing power is required such as high performance GPU which may perform parallel computing. The parallel computing is one of major requirement for deep learning. As the data set is huge and the applications may be combined with clusters or cloud computing, the high performance computing tools reduce the training time for a deep learning network. The architecture of deep learning can generalize the learning patterns and trends beyond immediate neighbors in non-local and global ways.
The deep learning can be applied in several applications such as voice control in phones, blue tooth speakers, intelligent TVs and in agriculture sectors. One of the most recent applications of this learning is driver-less cars where car itself recognize stop-signs or distinguish a pedestrian from a lamppost.
This chapter provides an overall development and concepts of deep learning along with the use of this technology in agriculture especially the identification of crop disease.
Chapter Objectives
The main objective of Deep Learning is known as a process to learn a structure of attributes at different levels. This process can be entirely unsupervised and tries to learn from the attributes of the previous levels to obtain and rebuild the original data.
The chapter is written for understanding the DL approaches. The objectives of this chapter are focusing on following two points
The readers will definitely be benefited through its unique content towards different sections such as Recurring Themes of DL. The recurring themes involves dynamic Programming for supervised and unsupervised learning.
The understanding of Convolutional Neural Network (CNN) is one of the important aspects in DL and have various features. Thus, the objective of the chapter is very clear and is written in structured format.
Author Contribution
Amit Sinha (AS) and Suneet Kr. Gupta (SKG) conceived of the presented idea. AS developed the theory and performed the computations. Dr. Amrita Chaturvedi (AC) encouraged AS and Anurag Tiwari (AT) to investigate and supervised the findings of this work followed by a verification of the analytical methods. All authors discussed the results and contributed to the final manuscript.
AC, AT and SKG carried out the experiment and wrote the manuscript with support from AS. SKG and AS developed the theoretical formalism, performed the analytic calculations and the numerical simulations. Both AS and AT contributed to the final version of the manuscript. All authors provided critical feedback and helped the research, analysis and writing the manuscript.