Classification of Leaf using Teachable Machine

Abstract

Leaf classification is an important stage in plant science. Manual classification of leaf is a quite difficult task for unknown person. For identification and classification of leaf type, machine learning (ML) based approach is simplest way. Henceforth, Google Teachable Machine (GTM) is an intuitive visual tool that provides workflow-oriented support for the development of machine learning models. This model helps to scan image object for leaf type identification. ML is a web-based tool allow user to train machine-learning model without any coding language. In the present work, Mango leaf, Orange leaf and Guava leaf has been selected for identification of its type. For the classification of leaves, hundred images of each type has been taken from available data set and categorized into three classes by giving Guava leaf, Orange leaf and Mango leaf name to created three classes. Teachable machine provides image, sound and pose training option for classification of an object. In this work, we used image option to train the model. This platform automatically trains the model using pre-trained deep learning algorithm. Leaf type verification is done by uploading new image of each leaf type that was not included in the training data set. After successful completion of experiment, result shows classification of each leaf type accurately.

Introduction

I. INTRODUCTION

Everyone see various types of plants in their day-to-day life. Some people have ability to classify it while some one cannot classify it properly [1]. Classification of plant or leaf type is quite difficult task for beginners. Generally, peoples believe on experience and dataset to classify an image, which is tedious, and time consuming work [2]. Use of technology makes us user friendly in carrying out image classification [3].

Image classification minimizes the gap between human perception and computer eyesight by training machine with available data set [4].  Modern computer with application of machine learning provide a deeper technical and economical farm management for identification of plant or leaf.

ML techniques show great potential in agriculture domain [5]. Artificial intelligence (AI) based machine learning (ML) algorithm is an excellent alternative for classification of leaf images. Prediction based on AI and ML becoming popular in research related to image classification [6].

It provides ability to learn automatically from experience. AI centers around the development of codes that can get to information and use it to find out on their own. The most common way of learning starts with perceptions or information, like models, direct insight, or guidance, to search for designs in information and pursue better choices later on in light of the models that we give. The main thing is to permit the PC learn automatically without human help and change actions accordingly [7]. The traditional approach of leaf classification needs prior knowledge of leaves.  New Approach avoids human errors and identifies the leaf accurately. It can exhibit smart intelligence, precise sensing along with good identification [8]. Google Teachable Machine (GTM) is a web-based GUI tool for creating custom machine learning classification models without any programming language and specialized technical expertise. It is developed to help students, teachers, designers, and others learn about ML by creating and using their own classification models.

In the present work, Google Teachable Machine based approach has been taken for classification of Guava leaf, Orange leaf and Mango leaf. The dataset for developing a well-trained machine-learning model has been taken from respective plant leaf available in surrounding.

Conclusion

Teachable Machine is not only an education tool, but a way for anyone to quickly and easily create a ML classifier using only their web browser in a matter of minutes. In this work, GTM is used for classification of leaves. It is the simplest way of classification of leaves without any programming skills. Figure 7 shows the classification accuracy of the proposed approach in comparison with other implementations.

References

[1] R. Sahay and R. Kaur, “Plant Recognition and its Care using TfLite, Teachable Machine”, Int. J. Res. Appl. Sci. & Engg., 11(5), 772-778,.(2023) [2] J. Jenn, N Wong and N. Fadzly, “Development of species recognition models using Google teachable machine on shorebirds and waterbirds”, J. Taibah Uni. Sci., 16(1), 1096-1111, (2022) [3] M. Aqil, F. Tabri, N. N. Andayani, S. Panikkai, Suwardi, E. Roy, Bunyamin Z, M. Azrai and T. Ratule, “Integration of smartphone technology for maize recognition”, IOP Conf. Series: Earth and Environmental Science 911 (2021) 012037 [4] R. Ahmad, Nasrullah, Vivek kumar, M. M. Tripathi and S. Singh, “Fault Detector using Image Classifier”, Int. J. Engg. Res. & Appl., 11(7),01-05, (2021) [5] T. Minche, V. Desai, M. Shaikh and R. Sonar, “Plantify-Disease Detector App”, Int. J. Res. Publication and Reviews, 4(6), 806-810, (2023) [6] R. Roy and A. K. Gupta, “Recognition of Distributed Combustion Regime From Deep Learning”, J. Energy Resources Tech., 144, 0923031-0923035, (2022) [7] O. Sri Nagesh, “Disease Prediction Using Various Machine Learning Algorithms”, Dickensian Journal, 22(4), 11-22, (2022) [8] C. Nesakumar, “Teachable Machine based Crop Health Monitoring System”, Bulletin Monumental, 22(9), 20-33, (2021) [9] https://teachablemachine.withgoogle.com/

Copyright

Copyright © 2023 R. K. Parate, K. M Dhole, S. J. Sharma. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.