Working Paper Article Version 1 This version is not peer-reviewed

Imbalance Data Set Classification of Tomato Pest Based on Lightweight CNN Model

Version 1 : Received: 20 July 2021 / Approved: 21 July 2021 / Online: 21 July 2021 (10:18:12 CEST)

How to cite: Liang, K.; Wang, Y.; Sun, L. Imbalance Data Set Classification of Tomato Pest Based on Lightweight CNN Model. Preprints 2021, 2021070481 Liang, K.; Wang, Y.; Sun, L. Imbalance Data Set Classification of Tomato Pest Based on Lightweight CNN Model. Preprints 2021, 2021070481

Abstract

In the classifying task of tomato pests, there are two most prominent problems: the difficulty in collecting balance data set of tomato pests and the difficulty in applying the mainstream CNN models that have too much parameter to application in handheld terminals. In this paper, from the perspective of balance data set and lightweight network framework. Firstly, we use data enhancement method to expand the collected imbalance tomato pest data set into 8 data sets with a total amount of 4,000, 4,800, 5,600, 6,400 according to the two standards of balance data set and imbalance data set. Moreover, a lightweight network model is established, which we named SSNet. Through comparative experiments with the above data sets, we discuss the following two questions respectively: 1. Influence of data set balance and total amount of data, verify the effect of the model. 2. The performance of SSNet in identification and classification of tomato pest. The experimental results show that the balance data set is better than the imbalance data set when the amount of data is small. When the total amount of data increases to a certain extent, the accuracy of model of the two kinds of data sets reaches the same continuously. The size of SSNet model constructed is only 0.398M. When the data set is balanced with 4,000, the optimizer is selected as Adam and the learning rate is 0.01, the accuracy can reach 98.8%, which is far higher than state of the art deep learning models and can better complete the classification and identification task of tomato pests.

Keywords

Tomato pests; CNN; Balance data set

Subject

MATHEMATICS & COMPUTER SCIENCE, Algebra & Number Theory

Comments (0)

We encourage comments and feedback from a broad range of readers. See criteria for comments and our diversity statement.

Leave a public comment
Send a private comment to the author(s)
Views 0
Downloads 0
Comments 0
Metrics 0


×
Alerts
Notify me about updates to this article or when a peer-reviewed version is published.
We use cookies on our website to ensure you get the best experience.
Read more about our cookies here.