In order to realize omnidirectional real-time detection and awareness in thermal infrared (IR) region,a kind of biomimetic miniature IR imaging detection system (BMIRIDS) has been successfully developed.It is classifi...In order to realize omnidirectional real-time detection and awareness in thermal infrared (IR) region,a kind of biomimetic miniature IR imaging detection system (BMIRIDS) has been successfully developed.It is classified as a long wave IR(LWIR) system and a mid-wave IR(MWIR) one.The former consists of a LWIR fish-eye lens and a LWIR focal plane array (FPA)-biomimetic electronic module,the latter consists of a MWIR fish-eye lens and a MWIR FPA-biomimetic electronic module.Both can provide complete hemispherical staring field of view,and have the capability to determine target bearings,to remove all stationary scene clutter and only detect moving targets,to sense targets range and their velocity.The capability closely approximates the vision process performed in nature by biological sensors.Furthermore,an additional biomimetic module block has been integrated into BMIRIDS.The module block imitating the function of human vision produces the logarithmic brightness image,and enhances the target image outline by virtue of the Roberts gradient difference.The logarithmic transform permits expanding the dynamic range of optical incidence without saturating the electric circuit,and makes the image more suitable for watching.DSP and FPGA perform the aforesaid vision functions that are advantageous to detect and track targets.Therefore,BMIRIDS can perform more powerful real-time signal and image processing on the resulting image data than the ordinary previous systems mainly by the aid of "software-only" processing.Because BMIRIDS has a 2π-steradian (sr.) staring field-of-view,two of such systems opposed to one another would provide a complete view of space around the sensor platform.In addition,its low power consumption makes itself meet the manifold platforms' needs.The practice has proved that BMIRIDS can truly realize "omnidirectional" "real-time" situation awareness and dynamic information acquirement.By virtue of biomimetic technology,some limitations of the current photoelectric equipment have been broken and a new type of IR equipment has been developed.展开更多
为解决人工对荧光原位杂交(Fluorescence In Situ Hybridization,FISH)荧光图像进行结果判读存在的效率低、劳动强度大等问题,针对FISH荧光图像细胞智能检测提出一种融合空域图像增强的改进YOLOv5算法。算法在原始YOLOv5神经网络模型基...为解决人工对荧光原位杂交(Fluorescence In Situ Hybridization,FISH)荧光图像进行结果判读存在的效率低、劳动强度大等问题,针对FISH荧光图像细胞智能检测提出一种融合空域图像增强的改进YOLOv5算法。算法在原始YOLOv5神经网络模型基础上,加入了空域图像增强模块,并选择了模块最佳增强系数,扩大了模型对荧光图像的对比度适应范围,提高了模型的特征提取能力和细胞检测准确率。实验结果显示,改进YOLOv5模型的平均精度均值(Mean Average Precision,mAP)为0.983,达到了比原始模型更优的训练效果和收敛速度,并且,改进YOLOv5模型的细胞识别率达到91.65%,比原始YOLOv5模型提升了9.19%。将细胞智能检测算法嵌入自主开发的荧光图像智能检测软件,结合荧光点检测算法,可给出有效判读结果。展开更多
目的:探讨多探针FISH技术在检测喀什地区维吾尔族急性白血病患者细胞遗传学异常中的应用效果。方法临床纳入新疆喀什地区第一人民医院和广东省第二人民医院2013年1月至2015年12月期间收治的100例急性白血病细胞遗传学异常患者,收集患...目的:探讨多探针FISH技术在检测喀什地区维吾尔族急性白血病患者细胞遗传学异常中的应用效果。方法临床纳入新疆喀什地区第一人民医院和广东省第二人民医院2013年1月至2015年12月期间收治的100例急性白血病细胞遗传学异常患者,收集患者的骨髓标本,均行预包被式多探针FISH技术与常规染色体核型分析技术检测,并进行对比分析;同时,对患者实施护理干预措施。结果100例急性白血病患者中,50例急性髓系白血病( AML)患者多探针FISH技术检测率60.0%明显高于常规染色体异常核型检测率28.0%,50例急性淋巴细胞白血病( ALL )患者多探针 FISH 技术检测率58.0%明显高于常规染色体异常核型检测率26.0%,差异均有统计学意义( P〈0.05)。结论喀什地区维吾尔族急性白血病患者细胞遗传学异常采用多探针FISH技术检测,具有省时、高效且准确率高等优点,值得临床上作为诊断急性白血病的重要手段推广应用。展开更多
基金supported by the National High-Tech Research & Development Program of China
文摘In order to realize omnidirectional real-time detection and awareness in thermal infrared (IR) region,a kind of biomimetic miniature IR imaging detection system (BMIRIDS) has been successfully developed.It is classified as a long wave IR(LWIR) system and a mid-wave IR(MWIR) one.The former consists of a LWIR fish-eye lens and a LWIR focal plane array (FPA)-biomimetic electronic module,the latter consists of a MWIR fish-eye lens and a MWIR FPA-biomimetic electronic module.Both can provide complete hemispherical staring field of view,and have the capability to determine target bearings,to remove all stationary scene clutter and only detect moving targets,to sense targets range and their velocity.The capability closely approximates the vision process performed in nature by biological sensors.Furthermore,an additional biomimetic module block has been integrated into BMIRIDS.The module block imitating the function of human vision produces the logarithmic brightness image,and enhances the target image outline by virtue of the Roberts gradient difference.The logarithmic transform permits expanding the dynamic range of optical incidence without saturating the electric circuit,and makes the image more suitable for watching.DSP and FPGA perform the aforesaid vision functions that are advantageous to detect and track targets.Therefore,BMIRIDS can perform more powerful real-time signal and image processing on the resulting image data than the ordinary previous systems mainly by the aid of "software-only" processing.Because BMIRIDS has a 2π-steradian (sr.) staring field-of-view,two of such systems opposed to one another would provide a complete view of space around the sensor platform.In addition,its low power consumption makes itself meet the manifold platforms' needs.The practice has proved that BMIRIDS can truly realize "omnidirectional" "real-time" situation awareness and dynamic information acquirement.By virtue of biomimetic technology,some limitations of the current photoelectric equipment have been broken and a new type of IR equipment has been developed.
文摘为解决人工对荧光原位杂交(Fluorescence In Situ Hybridization,FISH)荧光图像进行结果判读存在的效率低、劳动强度大等问题,针对FISH荧光图像细胞智能检测提出一种融合空域图像增强的改进YOLOv5算法。算法在原始YOLOv5神经网络模型基础上,加入了空域图像增强模块,并选择了模块最佳增强系数,扩大了模型对荧光图像的对比度适应范围,提高了模型的特征提取能力和细胞检测准确率。实验结果显示,改进YOLOv5模型的平均精度均值(Mean Average Precision,mAP)为0.983,达到了比原始模型更优的训练效果和收敛速度,并且,改进YOLOv5模型的细胞识别率达到91.65%,比原始YOLOv5模型提升了9.19%。将细胞智能检测算法嵌入自主开发的荧光图像智能检测软件,结合荧光点检测算法,可给出有效判读结果。
基金广东省科技计划项目(20150210)Fund program:Scientific and Technological Planning Project of Guangdong Province
文摘目的:探讨多探针FISH技术在检测喀什地区维吾尔族急性白血病患者细胞遗传学异常中的应用效果。方法临床纳入新疆喀什地区第一人民医院和广东省第二人民医院2013年1月至2015年12月期间收治的100例急性白血病细胞遗传学异常患者,收集患者的骨髓标本,均行预包被式多探针FISH技术与常规染色体核型分析技术检测,并进行对比分析;同时,对患者实施护理干预措施。结果100例急性白血病患者中,50例急性髓系白血病( AML)患者多探针FISH技术检测率60.0%明显高于常规染色体异常核型检测率28.0%,50例急性淋巴细胞白血病( ALL )患者多探针 FISH 技术检测率58.0%明显高于常规染色体异常核型检测率26.0%,差异均有统计学意义( P〈0.05)。结论喀什地区维吾尔族急性白血病患者细胞遗传学异常采用多探针FISH技术检测,具有省时、高效且准确率高等优点,值得临床上作为诊断急性白血病的重要手段推广应用。