技术
- 分析与建模 - 计算机视觉软件
- 分析与建模 - 机器学习
适用行业
- 农业
- 教育
适用功能
- 产品研发
- 质量保证
用例
- 计算机视觉
- 视觉质量检测
服务
- 测试与认证
- 培训
关于客户
本案例研究的客户是伦敦帝国理工学院的 BioMediA 研究小组和 iFind 项目。他们参与医学研究,并正在寻找方法来加快为医学领域的机器学习用例创建高质量训练数据的过程。他们与领先的机器学习应用训练数据平台 Labelbox 合作进行了一项实验,挑战了只有医学专家才能为临床深度学习模型提供高质量注释的假设。
挑战
为机器学习 (ML) 用例创建高质量的训练数据可能既昂贵又耗时,特别是对于需要领域专家审查和标记数据的专业领域。在医疗领域尤其如此,医生经常需要仔细标记或审查训练数据。这个过程可能非常艰巨且成本高昂,从而减慢了可能挽救生命的算法的开发速度。我们面临的挑战是找到一种方法来创建高质量的培训数据,同时减少医生的参与,从而降低成本并加快开发过程。
解决方案
伦敦帝国学院 BioMediA 研究小组和 iFind 项目的研究人员与 Labelbox 合作进行的一项研究挑战了只有医学专家才能为临床深度学习模型提供高质量注释的假设。该团队进行了一项实验,其中经过最低限度培训的新手标记人员的任务是注释胎儿超声图像以查找特定先天性心脏病的迹象。然后,研究人员在两个数据集上训练相同的算法:一个由 Labelbox 工作人员标记,另一个由医学图像专家标记。实验发现,这些新手注释者可以高标准执行复杂的医学图像分割任务,其标签质量与专家提供的标签相似。研究人员得出的结论是,当与现有方法结合使用来处理噪声注释并为模型选择信息最丰富的注释时,新手标记团队可以在专门医疗用例中开发高性能模型方面发挥至关重要的作用。
运营影响
数量效益
Case Study missing?
Start adding your own!
Register with your work email and create a new case study profile for your business.
相关案例.
![](/files/casestudy/Intelligent-Farming-with-ThingWorx-Analytics.png)
Case Study
Intelligent Farming with ThingWorx Analytics
Z Farms was facing three challenges: costly irrigation systems with water as a limited resource, narrow optimal ranges of soil moisture for growth with difficult maintenance and farm operators could not simply turn on irrigation systems like a faucet.
![](/files/casestudy/Greenhouse-Intelligent-Monitoring-and-Control-Solution.png)
Case Study
Greenhouse Intelligent Monitoring and Control Solution
Farming Orchids is the most successful form of precision farming in Taiwan, and also the most exported flower. Orchids need a specific temperature and humidity conditions to grow and bloom, and its flowering time may not be in line with market demands, so the price collapses when there is overproduction. Therefore, some farmers began to import automated greenhouse control systems for breeding and forcing, which not only improves quality, but also effectively controls the production period and yield to ensure revenue. In 2012, an orchid farmer built a Forcing Greenhouse of about 200 pings (approximately 661 Square Meters) in Tainan, Taiwan. The system integrator adopted Advantech’s APAX-5000 series programmable automation controllers to build the control platform, coupled with Advantech WebAccess HMI/SCADA software, to achieve cloud monitoring. The staff of the orchid field can monitor important data anytime via smart phone, iPad, and other handheld devices, and control the growth and flowering conditions. System requirements: In the past, most environmental control systems of orchid greenhouses in Taiwan used PLCs (Programmable Logic Controller) with poorscalability and control, and could not be connected to the Internet formonitoring from the cloud. For advanced database analysis and networking capability, the PC platform must be adopted. Therefore, PAC Systems (Programmable Automation Controller) with both PLC programming capabilities andPC functions is a better choice.The environmental control of the Orchid greenhouse switches on and off devices like fan, shade net, cooling/heat pump, liquid flow control, water-cooling wall etc. It is controlled by a control panel of electric controllers, and is driven by a motor, to adjust the greenhouse temperature, humidity, and other environmental conditions to the set parameters.
![](/files/casestudy/Precision-beekeeping-with-wireless-temperature-monitoring.png)
Case Study
Precision beekeeping with wireless temperature monitoring
Honeybees are insects of large economic value and provide a vital service to agriculture by pollinating a variety of crops. In addition, bees provide us with valuable products such as honey, beeswax, propolis, bee venom, etc. Monitoring of honeybee colony health, population, productivity, and environmental conditions affecting the colony health have always been exceedingly difficult tasks in apiculture. Research has shown that even small deviations (by more than 2°C) from the optimal temperatures have a significant influence on the development of the brood and the health of adult bees.
![](/files/casestudy/Enabling-Internet-of-Things-Innovation-in-Agriculture.png)
Case Study
Enabling Internet of Things Innovation in Agriculture
DigiBale, wanted to apply technology know-how and IP from implementations successfully to more agriculture sectors including cotton, forestry, sugarcane and cattle. However, farmers and growers still have worries about the connected technology.