项目名称: 基于部分参考图像质量评估的二维矢量图形快速渲染技术研究
项目编号: No.61300108
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 自动化技术、计算机技术
项目作者: 沈永珞
作者单位: 广东财经大学
项目金额: 23万元
中文摘要: 随着消费类电子产品爆炸式的增长,比如智能手机、数字地图导航设备的普及,二维矢量图形凭借其无限缩放后仍能保证图像不失真的优点,逐渐取代位图,用于实现高图像质量的人机交互界面。但是,在二维矢量图形渲染过程中,需处理经复杂计算后生成的大量图形几何开销,并且图形几何开销的多少严重的影响着渲染速度和渲染结果的图像质量。因此,在渲染过程中,需要一种评价几何开销和图像质量关系的方法,用于动态的指导几何开销的生成。然而,传统的全参考(或无参考)图像质量评估方法存在需要存储大量信息(或计算量巨大)的缺点。相比之下,部分参考图像评估方法则更适用于硬件资源有限的嵌入式设备。因此,本课题希望探索一种新的低存储量、低运算量的部分参考图像质量评估方法,建立关于二维矢量图形几何开销和图像质量的评价模型,并在前期开发完成的渲染引擎的架构上融合该模型,设计新的渲染引擎,实现在保证图像质量条件下的最优快速渲染。
中文关键词: 部分参考图像质量评估;开销质量评价模型;二维矢量图形;渲染;
英文摘要: With the increasing of consumer electronics, such as smart phones and navigators, user interfaces with high image quality can be achieved by two-dimensional (2D) vector graphics instead of bitmap images. Comparing with bitmap images, 2D vector graphics occupy less data information and achieve high image quality after arbitrary scaling or rotating. However, during the rendering process of 2D vector graphics, huge amount of geometric information which are generated by complex calculation heavily affect rendering speed and image quality. Hence, an evaluation system for balancing geometric cost and image quality is considered in the rendering engine. Moreover, the traditional methods such as full-reference or no-reference image quality assessments are not suitable for embedded system environment due to expensive memory usage or computation requirement. Hence, in this research, we hope to explore a low requirement image quality assessment method which is based on reduced-reference approach. And based on the developed rendering engine, we will build an evaluation system for balancing geometric cost and image quality to achieve a better performance with high image quality and fast rendering speed.
英文关键词: Reduced-reference image qualit;cost and image quality model;2D vector graphics;rendering;