Three main parameters were selected to study their importance in transformation by budmicroinjection in non-head Chinese cabbage [Brassica campestris ssp. chinensis (L.) Makinovar. communis Tsen et Lee]. The results s...Three main parameters were selected to study their importance in transformation by budmicroinjection in non-head Chinese cabbage [Brassica campestris ssp. chinensis (L.) Makinovar. communis Tsen et Lee]. The results showed that the developmental stage of floral bud, theconcentrations of sucrose and surfactant Silwet L-77 were critical for the successfultransformation by this new method. The suitable bud size is 2-3 mm in length, the favorableconcentration of sucrose and surfactant Silwet L-77 are 8 and 0.02% respectively. When thesucrose concentration was greater than 10% or that of Silwet L-77 was above 0.1%, the treatedbuds became yellow and finally blighted. 4/6 T1 seedlings resistant to kanamycin were positiveby PCR analysis, and T2 progeny of all these positive T1 plants have one or more hybridizingbands by Southern blot. Under 5% sucrose, 0.02% Silwet L-77 and grade 2 bud (2-3 mm in itslength) parameters, the most favorable transformation efficiency is about 0.56%, and meanefficiency reaches 0.16% in all experiments indicating that bud microinjection is potentialtransformation way in non-head Chinese cabbage.展开更多
构建大规模茶芽目标检测数据集是一项耗时且繁琐的任务,为了降低数据集构建成本,探索少量标注样本的算法尤为必要。本文提出了YSVD-Tea(YOLO singular value decomposition for tea bud detection)算法,通过将预训练模型中的基础卷积替...构建大规模茶芽目标检测数据集是一项耗时且繁琐的任务,为了降低数据集构建成本,探索少量标注样本的算法尤为必要。本文提出了YSVD-Tea(YOLO singular value decomposition for tea bud detection)算法,通过将预训练模型中的基础卷积替换为3个连续的矩阵结构,实现了对YOLOX算法结构的重构。通过维度变化和奇异值分解操作,将预训练权重转换为与重构算法结构相对应的权重,从而将需要进行迁移学习的权重和需要保留的权重分离开,实现保留预训练模型先验信息的目的。在3种不同数量的数据集上分别进行了训练和验证。在最小数量的1/3数据集上,YSVD-Tea算法相较于改进前的YOLOX算法,mAP提高20.3个百分点。对比测试集与训练集的性能指标,YSVD-Tea算法在测试集与训练集的mAP差距仅为21.9%,明显小于YOLOX的40.6%和Faster R-CNN的55.4%。在数量最大的数据集上,YOLOX算法精确率、召回率、F1值、mAP分别为86.4%、87.0%、86.7%和88.3%,相较于对比算法均最高。YSVD-Tea在保证良好性能的同时,能够更好地适应少量标注样本的茶芽目标检测任务。展开更多
文摘Three main parameters were selected to study their importance in transformation by budmicroinjection in non-head Chinese cabbage [Brassica campestris ssp. chinensis (L.) Makinovar. communis Tsen et Lee]. The results showed that the developmental stage of floral bud, theconcentrations of sucrose and surfactant Silwet L-77 were critical for the successfultransformation by this new method. The suitable bud size is 2-3 mm in length, the favorableconcentration of sucrose and surfactant Silwet L-77 are 8 and 0.02% respectively. When thesucrose concentration was greater than 10% or that of Silwet L-77 was above 0.1%, the treatedbuds became yellow and finally blighted. 4/6 T1 seedlings resistant to kanamycin were positiveby PCR analysis, and T2 progeny of all these positive T1 plants have one or more hybridizingbands by Southern blot. Under 5% sucrose, 0.02% Silwet L-77 and grade 2 bud (2-3 mm in itslength) parameters, the most favorable transformation efficiency is about 0.56%, and meanefficiency reaches 0.16% in all experiments indicating that bud microinjection is potentialtransformation way in non-head Chinese cabbage.
文摘构建大规模茶芽目标检测数据集是一项耗时且繁琐的任务,为了降低数据集构建成本,探索少量标注样本的算法尤为必要。本文提出了YSVD-Tea(YOLO singular value decomposition for tea bud detection)算法,通过将预训练模型中的基础卷积替换为3个连续的矩阵结构,实现了对YOLOX算法结构的重构。通过维度变化和奇异值分解操作,将预训练权重转换为与重构算法结构相对应的权重,从而将需要进行迁移学习的权重和需要保留的权重分离开,实现保留预训练模型先验信息的目的。在3种不同数量的数据集上分别进行了训练和验证。在最小数量的1/3数据集上,YSVD-Tea算法相较于改进前的YOLOX算法,mAP提高20.3个百分点。对比测试集与训练集的性能指标,YSVD-Tea算法在测试集与训练集的mAP差距仅为21.9%,明显小于YOLOX的40.6%和Faster R-CNN的55.4%。在数量最大的数据集上,YOLOX算法精确率、召回率、F1值、mAP分别为86.4%、87.0%、86.7%和88.3%,相较于对比算法均最高。YSVD-Tea在保证良好性能的同时,能够更好地适应少量标注样本的茶芽目标检测任务。
