Mobile Food Intake Visualization and Voice Recognize (FIVR)

描述（由申请人提供）： 不适当的饮食摄入评估工具阻碍了饮食与疾病之间关系的研究。适用于大型流行病学研究的方法（例如，饮食回忆、食物日记和食物频率调查表）具有相当大的不准确性，而更准确的方法（例如，代谢病房研究、双标记水）对于在基于人群的研究中的使用来说成本过高和/或劳动密集。需要一种简单、便宜、方便、有效的膳食测量工具，以更准确地确定人群的膳食摄入量。因此，我们提出了一个三阶段的项目来开发和测试一种新的评估工具，称为FIVR（食物摄入可视化和语音识别器），它使用了创新技术的新组合：先进的语音识别，可视化技术和自适应用户建模在电子系统中自动记录和评估食物摄入量。FIVR使用手机实时捕捉饮食摄入的录音和照片。这些双重数据源被发送到数据库服务器，用于通过语音识别和图像分析进行实时食物识别和份量测量的识别处理。用户模型将允许在图像本身可能无法产生准确结果的情况下增强对食物和制备方法的识别，因为系统通过经验学习并适应个人的食物模式。目标是将现有的语音和图像识别技术融合到一个系统中，该系统将通过食物类型和独特特征来识别食物，并通过从至少两个角度显示图像的电影剪辑来确定体积。所识别的项目将与食物成分数据库中的适当食物项目和数量相匹配，并计算营养摄入量。研究人员将能够通过饮食分析程序ProNutra的改编版本查看结果分析。拟议的方案将包括三个独立的阶段：1）技术开发、集成和测试; 2）控制饮食的有效性测试（代谢病房研究）; 3）真实世界效用测试。收集的数据的有效性将通过营养计算与代谢病房饮食的已知组成的匹配程度来判断。FIVR有可能建立一种高度准确的膳食摄入量评估方法，适用于人群水平的成本效益，从而推进关键的公共卫生目标。

DESCRIPTION (provided by applicant): Inadequate dietary intake assessment tools hamper studying relationships between diet and disease. Methods suitable for use in large epidemiologic studies (e.g., dietary recall, food diaries, and food frequency questionnaires) are subject to considerable inaccuracy, and more accurate methods (e.g., metabolic ward studies, doubly-labeled water) are prohibitively costly and/or labor-intensive for use in population-based studies. A simple, inexpensive and convenient, yet valid, dietary measurement tool is needed to provide more accurate determination of dietary intake in populations. We therefore propose a three-phase project to develop and test a new assessment tool called FIVR (Food Intake Visualization and Voice Recognizer) that uses a novel combination of innovative technologies: advanced voice recognition, visualization techniques, and adaptive user modeling in an electronic system to automatically record and evaluate food intake. FIVR uses cell phones to capture both voice recordings and photographs of dietary intake in real-time. These dual sources of data are sent to a database server for recognition processing for real-time food recognition and portion size measurement through speech recognition and image analysis. The user model will allow for enhanced identification of food and method of preparation in situations that images alone might not produce accurate results as the system learns through experience and adapts to the individual's food patterns. Objectives are to fuse existing voice and image recognition techniques into a system that will recognize foods by food type and unique characteristics and determine volume by film clip showing an image from at least two angles. The item identified will be matched to an appropriate food item and amount within a food composition database and nutrient intake computed. Researchers will be able to view the resulting analysis through an adapted version of the dietary analysis program, ProNutra. The proposed protocol will incorporate three discrete phases: 1) technology development, integration, and testing; 2) validity testing with a controlled diet (metabolic ward study); and 3) real-world utility testing. Validity of the data collected will be judged by how closely the nutrient calculations match the known composition of the metabolic ward diets consumed. FIVR has the potential to establish a method of highly accurate dietary intake assessment suitable for cost-effective use at the population level, and thereby advance crucial public health objectives.