毕业论文

打赏
当前位置: 毕业论文 > 计算机论文 >

基于光场相机的显著性检测方法研究

时间:2020-05-17 19:31来源:毕业论文
利用光场图像作为输入,采用DOG来检测Focusness特性,并通过位置线索获得前景似然得分(FLS)和背景似然得分(BLS),然后选择具有最高BLS的图像作为背景层,并用它来估计背景区域

摘要视觉显著性检测是计算机视觉领域中一直存在的问题。它的目的是找到图像中最吸引人类视觉注意的像素或者区域。本文探索的是把光场图像作为显著性检测输入的问题。分析了商业Lytro光场相机通过单次拍摄就可以捕捉场景的光场的可用性功能,以及其收集的光场数据特性,实现了一种基于光场的显著性检测方法,该方法利用光场图像作为输入,采用DOG来检测Focusness特性,并通过位置线索获得前景似然得分(FLS)和背景似然得分(BLS),然后选择具有最高BLS的图像作为背景层,并用它来估计背景区域。此外,选择具有较高FLS的图像作为前景的候选目标。最后,对全聚焦图像进行基于颜色对比度的显著性检测得到对比度显著图,并将其与检测到的前景候选目标带权相加得到最终的结果。从而开发出了一套专门针对于光场的全新的显著性检测算法,并且在采集到的光场数据集上与传统的显著性检测方法进行了比较。结果表明:该方法可以在一些极具挑战性的场景下可靠地完成显著性检测任务,如相似的前景和背景,杂乱的背景,复杂的闭塞等,并达到很高的精度和鲁棒性。49138

毕业论文关键词  视觉显著性  图像显著性区域检测  光场   Focusness

毕业设计说明书外文摘要

Title          Saliency Detection on Light Field                                                               

Abstract

Visual saliency detection has been a problem in the field of computer vision. Its purpose is to find the pixels or regions which attractive human visual attention of the image . This paper explores the problem of using light fields as input for saliency detection.And it also analysis the availability of commercial plenoptic cameras that capture the light field of a scene in a single shot and the light field database obtains by the commercial plenoptic camera.I further develop a new saliency detection algorithm tailored for light fields.The method using optical field as input,and it calculation the Focusness regions by DOG.Then compute a foreground likelihood score (FLS) and a background likelihood score (BLS) by measuring the focusness of pixels/regions. We select the layer with the highest BLS as the background and use it to estimate the background regions. In addition, we choose regions with a high FLS as candidate salient objects. Finally, we conduct contrast-based saliency detection on the all-focus image and combine its estimation with the detected foreground saliency candidates.For validation, we acquire a light field database of a range of indoor and outdoor scenes and generate the ground truth saliency map. And compare with other saliency detection algorithms  on the database .Experiments show that our saliency detection scheme can robustly handle challenging scenarios such as similar foreground and background, cluttered background, and images with multiple depth layers and with heavy occlusions, etc., and achieve high accuracy and robustness.

Keywords  Visual saliency   Salient region detection   Light field   Focusness

目   次

1 绪论 1

1.1 视觉显著性检测研究及其意义 1

1.2 视觉显著性检测国内外研究现状  2

1.3 基于光场的显著性检测及其意义  3

2 光场图像的处理 6

2.1  焦点图像和全聚焦图像  6

2.2  Focusness测量  8

2.3  背景的选择  9

2.4  前景的选择  11

3  基于颜色对比度的显著图 14

4  最终显著图的生成 16 基于光场相机的显著性检测方法研究:http://www.lwfree.com/jisuanji/lunwen_52053.html

------分隔线----------------------------
推荐内容