绵羊的姿态与其健康及福利密切相关。随着智能化畜牧业需求的增长,自动、准确地检测绵羊姿态尤为尤为重要。本研究提出基于Mask R-CNN基准网络的新型RDS-Mask R-CNN绵羊姿态检测算法,以Res2Net101作为特征提取网络,同时引入可变形卷积(D...绵羊的姿态与其健康及福利密切相关。随着智能化畜牧业需求的增长,自动、准确地检测绵羊姿态尤为尤为重要。本研究提出基于Mask R-CNN基准网络的新型RDS-Mask R-CNN绵羊姿态检测算法,以Res2Net101作为特征提取网络,同时引入可变形卷积(Deformable convolution network,DCN),以更精准捕捉绵羊在不同位置的姿态特征,并运用软非极大值抑制(Soft non-maximum suppression,Soft NMS)算法实现重叠实例目标的准确分割。结果表明:1)目标检测框架算法对比:与该领域最经典的YOLOv3和Faster R-CNN相比,改进的算法在平均精度均值(Mean average precision,mAP)上分别提升了16.68%和8.64%;2)不同改进策略的算法对比:改进算法相较于基准网络,边界框平均精度均值(Bounding box mean average precision,Bbox mAP)提高6.21%,分割平均精度均值(Segmentation mean average precision,Segm mAP)提高6.61%,分别达到87.34%和81.50%;3)相较于Mask R-CNN,改进模型在识别绵羊站立与躺卧姿态时边界框平均精度(Bounding box average precision,Bbox AP)分别提高了6.84%和5.58%,分割平均精度(Segmentation average precision,Segm AP)分别提高了7.25%和5.17%;4)模型可解释性可视化结果表明RDS-Mask R-CNN能精准捕获绵羊站立和躺卧姿态关键部位深度特征,表明模型自动检测可行且具有可解释性。综上,本研究提出的RDS-Mask R-CNN算法,有效提升了绵羊姿态检测的精准度,为智慧养殖提供了技术支撑。展开更多
Glaucoma is characterized by chronic progressive optic nerve damage and retinal ganglion cell death.Although extensive research has been conducted on neuroprotection for retinal ganglion cells,there is still no treatm...Glaucoma is characterized by chronic progressive optic nerve damage and retinal ganglion cell death.Although extensive research has been conducted on neuroprotection for retinal ganglion cells,there is still no treatment for clinical use.Recent evidence shows that extracellular vesicles isolated from a variety of stem cells are efficacious in retinal ganglion cell neuroprotection.In this study,we tested the novel extracellular vesicle source of the retinal progenitor R-28 cell line in vitro and in vivo.We isolated and characterized extracellular vesicles from R-28 cells and tested their therapeutic efficacy in terms of retinal ganglion cell survival in vitro and in an in vivo glaucoma model,measuring retinal ganglion cell survival and preservation of their axons.Additionally,we tested extracellular vesicles for their neuroprotective capacity in retinal ganglion cells differentiated from human embryonic stem cells.Finally,we investigated miRNA changes in retinal ganglion cells with R-28 extracellular vesicle treatment,and predicted possible pathways that may be modulated.R-28 extracellular vesicles improved retinal ganglion cell survival but failed to preserve axons significantly.Moreover,the results also illustrated the neuroprotection of R-28 extracellular vesicles on human retinal ganglion cells.Finally,we also showed changes in hsa-miRNA-4443,hsa-miRNA-216a-5p,hsa-let-7e-5p,hsa-miRNA-374b-5p,hsa-miRNA-331-3p,and hsa-miRNA-421 expressions,which may have neuroprotective potential on retinal ganglion cell degeneration.This study will pave the way for miRNA and extracellular vesicle-based neuroprotective therapies for glaucoma.展开更多
文摘绵羊的姿态与其健康及福利密切相关。随着智能化畜牧业需求的增长,自动、准确地检测绵羊姿态尤为尤为重要。本研究提出基于Mask R-CNN基准网络的新型RDS-Mask R-CNN绵羊姿态检测算法,以Res2Net101作为特征提取网络,同时引入可变形卷积(Deformable convolution network,DCN),以更精准捕捉绵羊在不同位置的姿态特征,并运用软非极大值抑制(Soft non-maximum suppression,Soft NMS)算法实现重叠实例目标的准确分割。结果表明:1)目标检测框架算法对比:与该领域最经典的YOLOv3和Faster R-CNN相比,改进的算法在平均精度均值(Mean average precision,mAP)上分别提升了16.68%和8.64%;2)不同改进策略的算法对比:改进算法相较于基准网络,边界框平均精度均值(Bounding box mean average precision,Bbox mAP)提高6.21%,分割平均精度均值(Segmentation mean average precision,Segm mAP)提高6.61%,分别达到87.34%和81.50%;3)相较于Mask R-CNN,改进模型在识别绵羊站立与躺卧姿态时边界框平均精度(Bounding box average precision,Bbox AP)分别提高了6.84%和5.58%,分割平均精度(Segmentation average precision,Segm AP)分别提高了7.25%和5.17%;4)模型可解释性可视化结果表明RDS-Mask R-CNN能精准捕获绵羊站立和躺卧姿态关键部位深度特征,表明模型自动检测可行且具有可解释性。综上,本研究提出的RDS-Mask R-CNN算法,有效提升了绵羊姿态检测的精准度,为智慧养殖提供了技术支撑。
基金supported by a Ph.D.scholarship from the YLSY program of the Republic of Turkiye,Ministry of National Educationfunded by Fight for Sight UK,grant reference#5183/5184。
文摘Glaucoma is characterized by chronic progressive optic nerve damage and retinal ganglion cell death.Although extensive research has been conducted on neuroprotection for retinal ganglion cells,there is still no treatment for clinical use.Recent evidence shows that extracellular vesicles isolated from a variety of stem cells are efficacious in retinal ganglion cell neuroprotection.In this study,we tested the novel extracellular vesicle source of the retinal progenitor R-28 cell line in vitro and in vivo.We isolated and characterized extracellular vesicles from R-28 cells and tested their therapeutic efficacy in terms of retinal ganglion cell survival in vitro and in an in vivo glaucoma model,measuring retinal ganglion cell survival and preservation of their axons.Additionally,we tested extracellular vesicles for their neuroprotective capacity in retinal ganglion cells differentiated from human embryonic stem cells.Finally,we investigated miRNA changes in retinal ganglion cells with R-28 extracellular vesicle treatment,and predicted possible pathways that may be modulated.R-28 extracellular vesicles improved retinal ganglion cell survival but failed to preserve axons significantly.Moreover,the results also illustrated the neuroprotection of R-28 extracellular vesicles on human retinal ganglion cells.Finally,we also showed changes in hsa-miRNA-4443,hsa-miRNA-216a-5p,hsa-let-7e-5p,hsa-miRNA-374b-5p,hsa-miRNA-331-3p,and hsa-miRNA-421 expressions,which may have neuroprotective potential on retinal ganglion cell degeneration.This study will pave the way for miRNA and extracellular vesicle-based neuroprotective therapies for glaucoma.
