EAGER: Advancing Digital Drawing by Analysis and Modeling of Design Intent

绘画是视觉思维和发现的无与伦比的辅助工具，被设计师、艺术家、科学家、工程师和技术人员所采用。虽然计算机以深刻的方式推动了对文本、摄影和音乐的处理，但在数字时代，绘画基本上保持不变。这个探索性的研究将包括一系列的研究，目的是在基本水平上理解绘画的过程，并着眼于在计算机上推进绘画作为一种图形媒体。项目成果将通过提供对绘图过程的基本见解，直接影响计算机在支持图形创意和跨广泛学科（从设计到工程到制造到教育）交流方面的作用。在这个项目中，机器学习方法将应用于一系列研究中的众包绘图任务，以产生对绘图过程和语义的新理解，最终目标是提升计算机上的绘图，使其不仅仅是模仿经典艺术媒体的尝试，而是提供不可或缺的新功能，使绘图作为一种视觉交流形式更具表现力和效率。为此，本工作将考虑以下问题。设计师绘制的笔触是什么？这些笔触是如何与所描绘的场景或概念联系起来的？画出的笔画如何被定性和量化？不同的设计师和他们所描绘的场景在绘图过程中有什么不同？有可能描述快速草图和完成的插图之间的区别吗？是否可以推断出绘图过程的模型，如果可以，这些模型是否可以用于理解和捕捉设计师的意图？可以开发一个定性形状（即随着绘图过程而演变的形状）的模型吗？该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Drawing is an incomparable aid to visual thinking and discovery, employed by designers, artists, scientists, engineers, and technologists alike. While computing has advanced the manipulation of text, photography, and music in profound ways, drawing has remained largely unchanged in the digital age. This exploratory research will involve a series of studies aimed at understanding the process of drawing at a basic level, with an eye to advancing drawing as a graphical medium on the computer. Project outcomes will have a direct impact on the role of computing in supporting graphical ideation and communication across a wide range of disciplines, from design to engineering to manufacturing to education, by providing fundamental insights into the drawing process. In this project machine learning methods will be applied to crowd-sourced drawing tasks across a series of studies to generate new understanding of the process and semantics of drawing, with the ultimate goal of elevating drawing on computers so that it is much more than an attempt to emulate classic art media but instead provides indispensable new capabilities that make drawing, as a form of visual communication, both more expressive and more efficient. To this end, this work will consider questions such as the following. What are the strokes that designers draw, and how do these strokes relate to the scene or concept that is depicted? How can the strokes drawn be characterized and quantified? How does the drawing process vary among individual designers and the scenes they depict? Is it possible to characterize the difference between a quick sketch and a finished illustration? Can models of the drawing process be inferred, and if so can such models be used for understanding and capturing a designer's intent? Can a model for qualitative shape (that is, shape that evolves with the drawing process) be developed?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.