ITR - (ASE+NHS) - (dmc+int): Triage and the Automated Annotation of Large Image Data Sets

提案：0427223主要研究人员：Donald Geman， Yali Amit， Stuart Geman和Laurent youn23机构：约翰霍普金斯大学，芝加哥大学，布朗大学和约翰霍普金斯大学提案标题：ITR-\(ASE+NHS)\-\(dmc+int)\：大型图像数据集的分类和自动注释摘要长期目标是在语义层面上为大型图像数据库的渐进式注释提供一个计算和数学框架。该研究将两种强大的范式（从粗到精的索引和句法场景解析）融合到一个模型和计算策略中，以提供作为可用计算周期函数的从少到多的场景详细注释。粗略的、可能有缺陷的解释出现在处理的早期阶段，随后是更精细、更准确的解释。从粗到精和句法模型之间的融合使人们能够获得两者的最佳效果：一个几乎最优的计算引擎，用于检测嵌入在完整语义和句法场景分析中的候选成分。大型图像数据集无处不在。来源包括医药、制造业、天体物理学、分子生物学、国防和情报。一般来说，这些资源的用处只与我们获取选定语义类别的能力成比例，比如“包括人”或“包含额叶肿块”。然后就非常需要自动注释，计算机程序将产生一个“元”数据结构，部分地描述每个图像的内容和上下文。自动化场景标注的进展将对包括医学和监控在内的广泛科学学科产生直接影响。

Proposal: 0427223Principal Investigators: Donald Geman, Yali Amit, Stuart Geman, and Laurent YounesInstitutions: Johns Hopkins U, U of Chicago, Brown U, and Johns Hopkins UProposal Title: ITR-\(ASE+NHS)\-\(dmc+int)\: Triage and Automated Annotation of large Image Data SetsABSTRACTThe long-term goal is a computational and mathematical framework for progressive annotation of large image databases at a semantic level. The proposed research fuses two powerful paradigms, coarse-to-fine indexing and syntactic scene parsing, into one model and computational strategy in order to provide a less-to-more detailed annotation of a scene as a function of available computing cycles. Coarse, and likely flawed, interpretations emerge at the early stages of processing, followed by finer and more accurate ones. The fusion between coarse-to-fine and syntactic models allows one to gain the best of both: a nearly optimal computational engine for detecting candidate constituents embedded in a full semantic and syntactic scene analysis.Large image data sets are ubiquitous. Sources include medicine, manufacturing, astrophysics, molecular biology, defense and intelligence. In general, these resources are useful only in proportion to our ability to access selected semantic categories, such as ``includes people'' or ``contains a frontal lobe mass''. There is then a great need for automated annotation, whereby computer programs would produce a ``meta'' data structure, partially describing the contents and context of each image. Progress in automated scene annotation will have an immediate impact on a broad range of scientific disciplines, including medicine and surveillance .