Web-SpikeSegNet: Deep Learning Framework for Recognition and Counting of Spikes From Visual Images of Wheat Plants

Tanuj Misra,Aditya Nigam,Rajni Jain,Rabi Narayan Sahoo,Mrinmoy Ray,Eldho Varghese,Sudhir Kumar,A R Rao,Sudeep Marwaha,Alka Arora,Shailendra Kumar,Viswanathan Chinnusamy,Ranjeet Ranjan Jha

doi:10.1109/access.2021.3080836

Abstract

Computer vision with deep learning is emerging as a significant approach for non-invasive and non-destructive plant phenotyping. Spikes are the reproductive organs of wheat plants. Detection and counting of spikes considered the grain-bearing organ have great importance in the phenomics study of large sets of germplasms. In the present study, we developed an online platform, “Web-SpikeSegNet,” based on a deep-learning framework for spike detection and counting from the wheat plant’s visual images. The architecture of the Web-SpikeSegNet consists of 2 layers. First Layer, Client-Side Interface Layer, deals with end user’s requests and corresponding responses management. In contrast, the second layer, Server Side Application Layer, consists of a spike detection and counting module. The backbone of the spike detection module comprises of deep encoder-decoder network with hourglass network for spike segmentation. The Spike counting module implements the “Analyze Particle” function of imageJ to count the number of spikes. For evaluating the performance of Web-SpikeSegNet, we acquired the wheat plant’s visual images, and the satisfactory segmentation performances were obtained as Type I error 0.00159, Type II error 0.0586, Accuracy 99.65%, Precision 99.59% and F1 score 99.65%. As spike detection and counting in wheat phenotyping are closely related to the yield, Web-SpikeSegNet is a significant step forward in the field of wheat phenotyping and will be very useful to the researchers and students working in the domain.

Highlights

Generic improvement in yield and climate resilience isWheat is one of the major food crops grown yearly 9 critical for sustainable food security
D_conv_3_2q pTranspose convolution operation followed by batch normalization and merge operation with the corresponding encoder block output qConvolution operation followed by batch normalization scription of each decoder block in the Decoder_SpikeSegNet is presented in the tabular form (Table 2) and the algorithm for implementing the Decoder_SpikeSegNet network is given in Algorithm 2.Bottleneck_SpikeSegNet network contains three hourglasses, which provide more confident segmentation by concentrating the essential features captured at various occlusions, scale, and view-points [8], [13]
E1 lies within 241 SpikeSegNet, a case study is presented here

Summary

Digital Object Identifier

Web-SpikeSegNet: deep learning framework for recognition and counting of spikes from visual images of wheat plants. TANUJ MISRA1,5, ALKA ARORA1, SUDEEP MARWAHA1,RANJEET RANJAN JHA2,MRINMOY RAY 1, RAJNI JAIN6, A R RAO 1,ELDHO VARGHESE 4, SHAILENDRA KUMAR 5, SUDHIR KUMAR3,ADITYA NIGAM 2,RABI NARAYAN SAHOO 3 AND VISWANATHAN CHINNUSAMY 3. The first author acknowledges the fellowship received from ICAR-IASRI, New Delhi, India, to undertake this research work as part of his Ph.D. program and Nanaji Deshmukh Plant Phenomics Facility, ICAR-IARI, New Delhi, for the facilities. Bhat from RLBCAU-Jhansi is highly acknowledged for assisting in the rebuttal preparation.

Generic improvement in yield and climate resilience is

Output feature size

RESULTS AND DISCUSSION