{"title":"GCPDFFNet:用于水稻爆炸识别的小物体检测","authors":"Dejin Xie, Wei Ye, Yining Pan, Jiaoyu Wang, Haiping Qiu, Hongkai Wang, Zhaoxing Li, Tianhao Chen","doi":"10.1094/PHYTO-09-23-0326-R","DOIUrl":null,"url":null,"abstract":"<p><p>Early detection of rice blast disease is pivotal to ensure rice yield. We collected in situ images of rice blast and constructed a rice blast dataset based on variations in lesion shape, size, and color. Given that rice blast lesions are small and typically exhibit round, oval, and fusiform shapes, we proposed a small object detection model named GCPDFFNet (global context-based parallel differentiation feature fusion network) for rice blast recognition. The GCPDFFNet model has three global context feature extraction modules and two parallel differentiation feature fusion modules. The global context modules are employed to focus on the lesion areas; the parallel differentiation feature fusion modules are used to enhance the recognition effect of small-sized lesions. In addition, we proposed the SCYLLA normalized Wasserstein distance loss function, specifically designed to accelerate model convergence and improve the detection accuracy of rice blast disease. Comparative experiments were conducted on the rice blast dataset to evaluate the performance of the model. The proposed GCPDFFNet model outperformed the baseline network CenterNet, with a significant increase in mean average precision from 83.6 to 95.4% on the rice blast test set while maintaining a satisfactory frames per second drop from 147.9 to 122.1. Our results suggest that the GCPDFFNet model can accurately detect in situ rice blast disease while ensuring the inference speed meets the real-time requirements.</p>","PeriodicalId":20410,"journal":{"name":"Phytopathology","volume":null,"pages":null},"PeriodicalIF":2.6000,"publicationDate":"2024-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"GCPDFFNet: Small Object Detection for Rice Blast Recognition.\",\"authors\":\"Dejin Xie, Wei Ye, Yining Pan, Jiaoyu Wang, Haiping Qiu, Hongkai Wang, Zhaoxing Li, Tianhao Chen\",\"doi\":\"10.1094/PHYTO-09-23-0326-R\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>Early detection of rice blast disease is pivotal to ensure rice yield. We collected in situ images of rice blast and constructed a rice blast dataset based on variations in lesion shape, size, and color. Given that rice blast lesions are small and typically exhibit round, oval, and fusiform shapes, we proposed a small object detection model named GCPDFFNet (global context-based parallel differentiation feature fusion network) for rice blast recognition. The GCPDFFNet model has three global context feature extraction modules and two parallel differentiation feature fusion modules. The global context modules are employed to focus on the lesion areas; the parallel differentiation feature fusion modules are used to enhance the recognition effect of small-sized lesions. In addition, we proposed the SCYLLA normalized Wasserstein distance loss function, specifically designed to accelerate model convergence and improve the detection accuracy of rice blast disease. Comparative experiments were conducted on the rice blast dataset to evaluate the performance of the model. The proposed GCPDFFNet model outperformed the baseline network CenterNet, with a significant increase in mean average precision from 83.6 to 95.4% on the rice blast test set while maintaining a satisfactory frames per second drop from 147.9 to 122.1. Our results suggest that the GCPDFFNet model can accurately detect in situ rice blast disease while ensuring the inference speed meets the real-time requirements.</p>\",\"PeriodicalId\":20410,\"journal\":{\"name\":\"Phytopathology\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":2.6000,\"publicationDate\":\"2024-07-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Phytopathology\",\"FirstCategoryId\":\"97\",\"ListUrlMain\":\"https://doi.org/10.1094/PHYTO-09-23-0326-R\",\"RegionNum\":2,\"RegionCategory\":\"农林科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"2024/7/5 0:00:00\",\"PubModel\":\"Epub\",\"JCR\":\"Q2\",\"JCRName\":\"PLANT SCIENCES\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Phytopathology","FirstCategoryId":"97","ListUrlMain":"https://doi.org/10.1094/PHYTO-09-23-0326-R","RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/7/5 0:00:00","PubModel":"Epub","JCR":"Q2","JCRName":"PLANT SCIENCES","Score":null,"Total":0}
GCPDFFNet: Small Object Detection for Rice Blast Recognition.
Early detection of rice blast disease is pivotal to ensure rice yield. We collected in situ images of rice blast and constructed a rice blast dataset based on variations in lesion shape, size, and color. Given that rice blast lesions are small and typically exhibit round, oval, and fusiform shapes, we proposed a small object detection model named GCPDFFNet (global context-based parallel differentiation feature fusion network) for rice blast recognition. The GCPDFFNet model has three global context feature extraction modules and two parallel differentiation feature fusion modules. The global context modules are employed to focus on the lesion areas; the parallel differentiation feature fusion modules are used to enhance the recognition effect of small-sized lesions. In addition, we proposed the SCYLLA normalized Wasserstein distance loss function, specifically designed to accelerate model convergence and improve the detection accuracy of rice blast disease. Comparative experiments were conducted on the rice blast dataset to evaluate the performance of the model. The proposed GCPDFFNet model outperformed the baseline network CenterNet, with a significant increase in mean average precision from 83.6 to 95.4% on the rice blast test set while maintaining a satisfactory frames per second drop from 147.9 to 122.1. Our results suggest that the GCPDFFNet model can accurately detect in situ rice blast disease while ensuring the inference speed meets the real-time requirements.
期刊介绍:
Phytopathology publishes articles on fundamental research that advances understanding of the nature of plant diseases, the agents that cause them, their spread, the losses they cause, and measures that can be used to control them. Phytopathology considers manuscripts covering all aspects of plant diseases including bacteriology, host-parasite biochemistry and cell biology, biological control, disease control and pest management, description of new pathogen species description of new pathogen species, ecology and population biology, epidemiology, disease etiology, host genetics and resistance, mycology, nematology, plant stress and abiotic disorders, postharvest pathology and mycotoxins, and virology. Papers dealing mainly with taxonomy, such as descriptions of new plant pathogen taxa are acceptable if they include plant disease research results such as pathogenicity, host range, etc. Taxonomic papers that focus on classification, identification, and nomenclature below the subspecies level may also be submitted to Phytopathology.