MRI: Acquisition of OCT Imaging System and Deep Learning Workstation for Interdisciplinary Healthcare Research and Education

该提案要求在加州州立大学（CSU，Chico）采购扫频源光学相干断层扫描（OCT）成像系统和深度学习工作站（DLW）。拟议的OCT成像系统和DLW将使OCT成像在牙科中的应用研究能够检测早期龋损、微骨折、牙髓炎症、口腔恶性肿瘤的早期发育不良变化和其他牙科疾病的特征。OCT成像系统和DLW的收购将显着增强当前的研究计划，并使新的研究方向在CSU，奇科。这些仪器还将支持与斯托尼布鲁克大学牙科医学院教师的既定合作研究。来自电气和计算机工程，计算机科学，机械和机电一体化工程和可持续制造部门的约500名学生将成为OCT系统和DLW的主要用户，并将可供更广泛的CSU，奇科研究社区使用。 CSU，奇科是一个少数民族服务的机构，有很大比例的学生来自代表性不足和服务不足的群体，包括退伍军人和西班牙裔。这种先进的OCT系统和DLW将为这些学生提供实践经验。OCT系统和DLW将用于（1）增加本科生参与研究，（2）促进主动学习和技能发展，（3）培训高年级本科生和研究生最先进的成像和算法开发技术，以及（4）刺激与CSU，Chico和其他高等院校的教师和学生的合作。 拟议的扫频源光学相干断层扫描（OCT）成像系统和深度学习工作站（DLW）将刺激芝加哥州立大学医疗保健和工业无损检测的跨学科研究项目。OCT是一种基于低相干干涉测量的非侵入性光学成像模式，其利用非电离近红外激光获得1 - 10微米分辨率的图像。目前，OCT的主要生物医学应用是在眼科。OCT的许多其他应用正在研究中，因为研究人员利用其快速非侵入性获取图像的能力。机器学习和深度学习技术可用于补充OCT图像，以更准确地识别患病和受损组织。CSU Chico的以下研究项目将利用OCT成像系统和DLW：（1）开发深度学习模型，即卷积神经网络（CNN），并对OCT图像进行定量分析以进行早期龋齿检测;（2）研究各种深度学习优化方法及其与OCT图像的性能，以最大限度地减少反向传播误差;（3）开发信号和图像处理算法，从OCT数据中提取有意义的特征用于图像分类：（4）设计紧凑、低成本的光纤探头，用于咬合面龋的非侵入性OCT成像;以及（5）使用深度学习来分析和建模非语音声音，例如音乐或工业噪声，以及音乐音色的自动分析和分类。OCT成像系统和DLW将有助于促进工程，科学和农业教师之间的跨学科努力，并在CSU，Chico开发本科生和研究生研究和教学实验室。该奖项反映了NSF的法定使命，并被认为值得通过使用基金会的智力价值和更广泛的影响审查标准进行评估来支持。

This proposal requests acquisition of a swept-source optical coherence tomography (OCT) imaging system and a deep learning workstation (DLW) at California State University, Chico (CSU, Chico). The proposed OCT imaging system and DLW will enable studies on the application of OCT imaging in dentistry to detect early carious lesions, micro-fractures, pulpal inflammation, early dysplastic changes in oral malignancies and signatures of other dental diseases. Acquisition of the OCT imaging system and DLW will significantly enhance current research programs and enable new research directions at CSU, Chico. These instruments will also support the established collaborative research with faculty at the Stony Brook University School of Dental Medicine. Approximately 500 students from the Departments of Electrical and Computer Engineering, Computer Science, and Mechanical and Mechatronic Engineering and Sustainable Manufacturing will be among the major users of the OCT system and DLW and will be available for use by the broader CSU, Chico research community. CSU, Chico is a minority-serving institution with a large proportion of students from underrepresented and underserved groups, including veterans, and Hispanics. This cutting-edge OCT system and DLW will provide these students with hands-on experience. The OCT system and DLW will be used to (1) increase involvement of undergraduate students in research, (2) promote active learning and skills development, (3) train upper-division undergraduate and graduate students on state-of-the-art imaging and algorithm development techniques, and (4) stimulate collaborations with faculty and students across CSU, Chico and from other institutions of higher learning. The proposed swept-source optical coherence tomography (OCT) imaging system and deep learning workstation (DLW) will stimulate interdisciplinary research projects in healthcare and industrial nondestructive testing at CSU, Chico. OCT is a noninvasive optical imaging modality based on low-coherence interferometry that utilizes non-ionizing near-infrared laser to obtain images with 1-10 micrometer resolution. Currently, the major biomedical application of OCT is in ophthalmology. Many other applications of OCT are under investigation as researchers take advantage of the ability to rapidly acquire images noninvasively. Machine learning and deep learning techniques can be used to supplement the OCT images to more accurately identify diseased and damaged tissue. The following research projects at CSU, Chico will utilize the OCT imaging system and DLW: (1) the development of a deep learning model, namely convolutional neural networks (CNN), and quantitative analysis of OCT images for early dental caries detection; (2) the investigation of various deep learning optimization methods and their performances with OCT images to minimize back-propagating errors; (3) the development of signal and image processing algorithms to extract meaningful features from OCT data for image classification; (4) the design of compact and low-cost fiber optic based probe for noninvasive OCT imaging of occlusal caries; and (5) the use of deep learning to analyze and model non-speech sounds, such as music or industrial noise, as well as automatic analysis and classification of musical timbre. The OCT imaging system and DLW will help catalyze interdisciplinary efforts between engineering, science, and agriculture faculty, and develop an undergraduate and graduate research and teaching laboratory at CSU, Chico.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.