Table 2. Comparison of the proposed model with other related studies.

Researchers	Methods	Dataset (own or publicly available)	Camera to capture data	Number of observation	Learning rate	Number of iteration	Performance (%)
Zhou et al. (2019)	FCM-KM and Faster R-CNN fusion	Rice field of the Hunan Rice Research Institute, China	Canon EOS R (pixel: 2,400 * 1,600)	3,010	0.001	15,000	Rice blast: 96.71 Bacterial blight: 97.53 Blight: 98.26
Sethy et al. (2020)	Faster R-CNN	Farm field	Smartphone camera (48 Megapixel)	50	0.001	5	Initial steps to make a prototype for automatic detection of RFS Rice false smut
Phadikar, Sil & Das (2012)	Bayes’ and SVM Classifier	Rice field images of East Midnapur, India	Nikon COOLPIX P4 digital camera	1,000	–	–	Normal leaf image: 92 Brown spot image: 96.4 Blast image: 84 Bayes’ classifier: 79.5 SVM: 68.1
Ramesh & Vydeki (2020)	Optimized Deep Neural Network with Jaya Optimization Algorithm (DNN_JOA)	Farm field	High resolution digital camera	650	–	–	Rice blast: 98.9 Bacterial blight: 95.78 Sheath rot: 92 Brown spot: 94 Normal leaf: 90.57
Li et al. (2020)	Faster-RCNN	Rice field in Anhui, Jiangxi and Hunan Province, China	Mobile phone camera (iPhone7 & HUAWEI P10) and Sony DSC-QX10 camera	5,320	0.002	50,000	Rice sheath blight: 90.9 Rice stem borer: 71.4 Rice brown spot: 90
Prajapati, Shah & Dabhi (2017)	SVM	Farm field	NIKON D90 digital SLR (12.3 megapixels)	120	–	–	For SVM: 93.33 (training) 73.33 (testing) 5-fold cross-validation: 83.80 10-fold cross-validation: 88.57
Narendra Pal Singh Rathore (2020)	CNN	Kaggle dataset	–	1,000	–	–	Prediction accuracy: 99.61 (healthy and leaf blast)
Rahman et al. (2020)	Simple CNN	Rice fields of Bangladesh Rice Research Institute (BRRI)	Four different types of camera	1,426	0.0001	100	Mean validation accuracy: 94.33
Proposed Model	Faster-RCNN	Both on field data and Kaggle dataset	Smartphone camera (Xiaomi Redmi 8)	16,800	0.0002	50965	Rice blast: 98.09 Brown spot:98.85 Hispa: 99.17 Healthy rice leaf: 99.25