Plant Recognition using Spatial Transformer Network
Azhar Talha Syed1, Suresh Merugu2, Vijaya Kumar Koppula3

Manuscript received on 07 February 2019 | Revised Manuscript received on 20 February 2019 | Manuscript Published on 04 March 2019 | PP: 334-336 | Volume-7 Issue-5S2 January 2019 | Retrieval Number: ES2058017519/19©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (

Abstract: Agriculture is one of the most prominent work sectors in countries like India. However, the majority of farmers are unaware of the modern plant diseases and the methods are to be followed to expect a better yield from their crops. Data science and Machine Learning have made a great progress in recent years for providing a solution to problems like these. Findings: By developing a system which will help the farmers in getting aware about the different species of plants without having a need for definite education would be very helpful to them. Objective: In this paper, we propose an efficient way of recognizing plants using cell phone cameras, as it will be very easy for the farmers and also other people who have their work involving plants, to get information about a plant which will help them in their work. We also provide a performance analysis on our solution and the previous work in this paper. Methods/Statistical Analysis: In Machine Learning terminology this is a multiclass classification problem where the input is an image and the expected output is the class of which the plant in the image belongs to. There are several ways of solving a multi-class classification problem such as using K nearest neighbors, Multiclass Support Vector Machines, Neural Networks, and Convolutional Neural Networks. But for this problem, we also take user convenience into consideration and we suggest the use of Spatial Transformer Network as the classification will still be accurate whilst the image is not properly aligned and has a lot of noise in it.
Keywords: Plant Recognition; Deep Learning; Convolutional Neural Networks; Spatial Transformer Network.
Scope of the Article: Pattern Recognition