Algorithm Development For Reconstruction Of Design Elements

设计元素重构的算法开发

• 批准号: 1658987
• 负责人: Karen Smith
• 金额: $ 21.3万
• 依托单位: University of South Carolina at Columbia
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-05-01 至 2020-04-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1658987&HistoricalAwards=false
• 关键词: Algorithm; Development; Reconstruction; Design; Elements

Archaeologists, forensic scientists, and air crash examiners often find objects of evidence broken in many pieces, and if the number is large they face the problem of how to match, edge shapes and surface markings to fit the fragments back together. In cases where individual sites often yield thousands of pieces this problem can be particularly acute. This research will develop a computer assisted program, "SnowVision" to aid in fitting fragments of objects together. The research is of significance for its use not only at archaeological sites but also for the general applicability of the algorithms developed. The project will provide new insights into a novel line of research for computer vision scientists. The open-source nature of the project means that the products delivered by the team can be addressed to multiple challenges. In addition to the educational impacts, the project includes outreach efforts to Native-American tribes and institutions.Dr. Karen Smith and her research team are addressing this issue by focusing on a specific type of prehistoric Southeastern US pottery which contains paddle-stamped designs. For the development and implementation of SnowVision, Dr. Smith and her team focus on the designs embellished on Swift Creek pottery, ca. AD 200-800, found on sites throughout the Deep South. Swift Creek's curvilinear designs, which were carved onto wooden paddles and subsequently stamped onto pottery vessels, offer insights into networks of exchange, ritual, and mobility at an unprecedented level of granularity when paddle matches are found across archaeological contexts, as they so often are. Limitations of time, design expertise, and collections access have been the primary impediments to advancement in the study of these pre-Columbian networks, for which design matches provide unequivocal evidence. SnowVision will increase the speed by which researchers are able to identify design matches from different archaeological sites. This will, in turn, incentivize the digitization of understudied collections. By making the software and related pottery sherd images available online, SnowVision also will increase the reach of collections curated in different museums. Dissemination of the designs will continue to raise awareness of the deep decorative traditions of our Nation's indigenous communities. Sherd images will be archived with Open Context, where sherd data can be leveraged with data from other research projects.

考古学家、法医学家和空难调查员经常发现证据物品破碎成许多碎片，如果数量很大，他们面临如何匹配、边缘形状和表面标记的问题，以将碎片重新组合在一起。在个别网站经常产生数千件的情况下，这个问题可能特别严重。这项研究将开发一个计算机辅助程序“SnowVision”，以帮助将物体的碎片拼接在一起。 该研究不仅在考古遗址中使用，而且对于所开发的算法的普遍适用性具有重要意义。该项目将为计算机视觉科学家提供新的研究方向。该项目的开源性质意味着团队交付的产品可以应对多种挑战。除了教育方面的影响，该项目还包括对美国原住民部落和机构的推广工作。凯伦·史密斯博士和她的研究团队正在通过关注美国东南部一种特定类型的史前陶器来解决这个问题，这种陶器包含桨印设计。对于SnowVision的开发和实施，史密斯博士和她的团队专注于斯威夫特溪陶器上的设计。公元200-800年，在整个南方腹地的遗址上发现。斯威夫特克里克的曲线设计，这是雕刻在木制桨，随后印在陶器上，提供了深入了解交换网络，仪式，和流动性在一个前所未有的粒度水平时，桨比赛发现在考古背景下，因为他们经常是。时间、设计专业知识和藏品获取的限制一直是这些前哥伦布时期网络研究进展的主要障碍，设计匹配为这些网络提供了明确的证据。SnowVision将提高研究人员识别不同考古遗址设计匹配的速度。这将反过来激励未充分研究的收藏品的数字化。通过在线提供软件和相关的陶器碎片图像，SnowVision还将增加不同博物馆收藏的覆盖面。设计的传播将继续提高对我国土著社区深厚装饰传统的认识。Sherd图像将与Open Context一起存档，其中Sherd数据可以与其他研究项目的数据一起使用。

项目成果

A Framework for Design Identification on Heritage Objects

• DOI: 10.1145/3332186.3332190
• 发表时间: 2019-07
• 期刊: Proceedings of the Practice and Experience in Advanced Research Computing on Rise of the Machines (learning)
• 作者: Jun Zhou;Yuhang Lu;Karen Smith;Colin Wilder;Song Wang;P. Sagona;Ben Torkian

Building Science Gateways for Humanities

构建人文科学门户

• DOI: 10.1145/3311790.3396628
• 发表时间: 2020
• 期刊: PEARC '20: Practice and Experience in Advanced Research Computing
• 作者: Zhou, Jun;Smith, Karen;Wilsbacher, Greg;Sagona, Paul;Reddy, David;Torkian, Ben
• 通讯作者: Torkian, Ben

Curve-Structure Segmentation from Depth Maps: A CNN-based Approach and Its Application to Exploring Cultural Heritage Objects

• DOI: 10.1609/aaai.v32i1.12306
• 发表时间: 2017-11
• 作者: Yuhang Lu;Jun Zhou;Jing Wang;Jun Chen;Karen Smith;Colin Wilder;Song Wang