TY - JOUR

T1 - Stomatal State Identification and Classification in Quinoa Microscopic Imprints through Deep Learning

AU - Razzaq, Abdul

AU - Shahid, Sharaiz

AU - Akram, Muhammad

AU - Ashraf, Muhammad

AU - Iqbal, Shahid

AU - Hussain, Aamir

AU - Zia, M. Azam

AU - Qadri, Sulman

AU - Saher, Najia

AU - Shahzad, Faisal

AU - Shah, Ali Nawaz

AU - Rehman, Aziz-ur

AU - Jacobsen, Sven-Erik

PY - 2021

Y1 - 2021

N2 - Stomata are the main medium of plants for the trade of water, regulate the gas exchange, and are responsible for the process of photosynthesis and transpiration. The stomata are surrounded by guard cells, which help to control the rate of transpiration by opening and closing the stomata. The stomata states (open and close) play a significant role in describing the plant's health. Moreover, stomata counting is important for scientists to investigate the numbers of stomata that are open and those that are closed to measure their density and distribution on the surface of leaves through different sampling techniques. Although a few techniques for stomata counting have been proposed, these approaches do not identify and classify the stomata based on their states in leaves. In this research, we have developed an automatic system for stomata state identification and counting in quinoa leaf images through the transformed learning (neural network model Single Shot Detector) approach. In leaf imprint, the state of stomata has been determined by measuring the correlation between the area of stomata and the aperture of each detected stoma in the image. The stomata states have been classified through the Support Vector Machine (SVM) algorithm. The overall identification and classification accuracy of the proposed system are 98.6% and 97%, respectively, helping researchers to obtain accurate stomatal state information for leaves in an efficient and simple way.

AB - Stomata are the main medium of plants for the trade of water, regulate the gas exchange, and are responsible for the process of photosynthesis and transpiration. The stomata are surrounded by guard cells, which help to control the rate of transpiration by opening and closing the stomata. The stomata states (open and close) play a significant role in describing the plant's health. Moreover, stomata counting is important for scientists to investigate the numbers of stomata that are open and those that are closed to measure their density and distribution on the surface of leaves through different sampling techniques. Although a few techniques for stomata counting have been proposed, these approaches do not identify and classify the stomata based on their states in leaves. In this research, we have developed an automatic system for stomata state identification and counting in quinoa leaf images through the transformed learning (neural network model Single Shot Detector) approach. In leaf imprint, the state of stomata has been determined by measuring the correlation between the area of stomata and the aperture of each detected stoma in the image. The stomata states have been classified through the Support Vector Machine (SVM) algorithm. The overall identification and classification accuracy of the proposed system are 98.6% and 97%, respectively, helping researchers to obtain accurate stomatal state information for leaves in an efficient and simple way.

U2 - 10.1155/2021/9938013

DO - 10.1155/2021/9938013

M3 - Journal article

VL - 2021

JO - Complexity (Print)

JF - Complexity (Print)

SN - 1076-2787

M1 - 9938013

ER -