基于二维手绘草图的三维形状检索

随着互联网上三维模型数量的迅猛增多,需要发展便捷、可靠的基于内容的三维形状(模型)检索引擎。手绘草图具有以视觉形象表达概念的能力,符合人类传递信息的习惯,因而成为三维形状检索的重要交互手段。然而,直接表达人类意识的二维手绘草图与标准的三维形状表达之间常常存在较大的语义鸿沟,这给基于手绘草图的三维形状检索带来很多技术难题。文章对基于手绘草图的三维形状检索相关背景以及技术进展进行了介绍,重点介绍了用上下文信息填补语义鸿沟的最新方法。

Three Dimensional Shape Retrieval Based on Two Dimensional Skeches

With the rapid increase of 3D model on the internet, it is necessary to develop a convenient and reliable content-based 3D shape (model) search engine. Sketch has the ability to express concept as a visual icon, which is in line with the human habit of passing information, and thus becomes an important interactive means in 3D shape retrieval. However, there is often a larger semantic gap between the standard 3D shape representation and the 2D sketch which expresses human consciousness, which brings a lot of technical challenges to the sketch-based shape retrieval. In this paper, the relevant background and technological advances on the sketch-based 3D shape retrieval were introduced, and the latest methods that use context information to fill the semantic gap were particularly focused on.