基于OpenGL的交互式全景视频显示系统

随着虚拟现实和多媒体通信技术的快速发展,全景视频作为一种新兴的数字媒体,凭借其良好的临场感和沉浸感在影视娱乐、教育教学、工业仿真、医学研究等众多领域取得广泛应用。为了更好地开展全景视频的相关研究,建立交互式的全景视频显示系统成为了优化和完善全景视频业务的重要前提。本文基于OpenGL渲染技术,结合用户对于全景视频交互操作的需求,搭建了交互式全景视频显示系统,并在该系统上通过实验分析了本系统的交互时延和视频分辨率、帧率的关系。实验结果表明,交互产生的时延和视频分辨率、帧率的大小呈正相关关系。在本平台的设置下,交互产生的时延不超过0.02s,为用户提供了良好的交互性和视觉体验,为全景视频的相关研究与应用开发提供了稳定的平台。     

基于人眼检测优化的立体人机交互技术的研究

随着3D显示技术的逐渐成熟,越来越多基于3D显示技术的产品被应用于各个领域,例如3D电影、裸眼3D电视、体感交互式系统和影像教学等。基于3D显示的人机交互系统也是近年来越来越热门的研究方向,但是却没有高效地利用深度信息作为交互的线索,因此本文提出将3D显示与体感交互结合,实现立体人机交互的虚拟键盘,让用户只需点击屏幕前方键盘的图像即可进行对应按键信息的录入和编辑,其帧率可达10.6fps。针对3D显示中用户视角偏离会导致按键的像在视觉上的偏移,容易造成误输入的问题,在系统中整合人眼检测功能,修正因视角偏离带来按键定位误差,从而实现了准确率100%的字符输入交互功能。     

基于S曲线全局动态调光的改进算法

S曲线全局动态调光算法可以降低LED液晶显示器的功耗,同时能够提高显示图像的静态对比度,但该算法会造成部分图像色彩失真和细节丢失。针对这一问题,本文提出一种图像细节层分离与视觉显著性理论相结合的S曲线改进算法。首先,将原始图像转换至HSV色彩空间进行亮度和色度分离;然后,在图像亮度分量上采用双边滤波得到图像的基础层与细节层,基础层采用S曲线进行动态范围拉伸,实现像素补偿,细节层则运用视觉显著性理论进行分区与权值增强,弥补由像素补偿带来的细节损失;最后,将处理后的各层图像转换至RGB空间显示。将本文算法的仿真结果与原S曲线算法的结果进行对比。结果显示,本文算法在维持原算法功耗降低和静态对比度提升水平不变的基础上,解决了原算法在部分图像中出现的色彩失真和细节丢失问题,提升了图像的视觉显示效果,同时本文算法的仿真结果具有更大的信息熵和平均梯度。     

基于显著区域提取的红外图像舰船目标检测

针对复杂海面背景下红外图像舰船目标由于灰度不均匀、海杂波干扰大等因素造成的自动检测虚警率高、准确率低的问题,提出了一种显著区域提取和目标精确分割相结合的红外舰船目标检测方法。首先,利用基于图论的视觉显著性(Graph-based Visual Saliency ,GBVS)模型计算待检测图像的显著图,使得目标区域信息增强;其次,结合舰船目标先验信息（长短轴、面积等）,利用多级阈值划分算法提取关注的显著区域,并确定原图中候选目标区域;最后,利用空间约束模糊C均值(Fuzzy C-Means,FCM)算法对候选区域进行分割,结合目标先验知识对分割区域筛选并输出目标位置。所提方法在公开数据集IRShips上与相关方法进行比较,结果表明,相比直接进行全图目标搜索的方法,所提方法不仅准确率高、执行速度快,且检测目标的位置更加精确。     

Exploiting object features in deep gaze prediction models

The human visual system analyzes the complex scenes rapidly. It devotes the limited perceptual resources to the most salient subsets and/or objects of scenes while ignoring their less salient parts. Gaze prediction models try to predict the human eye fixations (human gaze) under free-viewing conditions while imitating the attentive mechanism. Previous studies on saliency benchmark datasets have shown that visual attention is affected by the salient objects of the scenes and their features. These features include the identity, the location, and the visual features of objects in the scenes, beside to the context of the input image. Moreover, the human eye fixations often converge to the specific parts of salient objects in the scenes. In this paper, we propose a deep gaze prediction model using object detection via image segmentation. It uses some deep neural modules to find the identity, location, and visual features of the salient objects in the scenes. In addition, we introduce a deep module to capture the prior bias of human eye fixations. To evaluate our model, several challenging saliency benchmark datasets are used in the experiments. We also conduct an ablation study to show the effectiveness of our proposed modules and its architecture. Despite its fewer parameters, our model has comparable, or even better performance on some datasets, to the state-of-the-art saliency models.     

Perceptual visual quality assessment using deeply-learned gaze shifting kernel

Image quality assessment (IQA) is a useful technique in computer vision and machine intelligence. It is widely applied in image retrieval, image clustering and image recognition. IQA algorithms generally rely on human visual system (HVS), which can reflect how human perceive salient regions in the image. In this paper, we leverage both low-level features and high-level semantic features to select salient regions, which will be concatenated to form GSPs by the designed saliency-constraint algorithm to mimic human visual system. We design an enhanced IQA index based on the GSPs to calculate the simialrity between reference image and test image to achieve image quality assessment. Experiments demonstrate that our IQA method can achieve satisfactory performance.     

基于无监督学习的单人多姿态图像生成方法

针对目前视觉监控领域中采集到的人物数据样本量少和特征单一的问题,提出了一种具有高视觉感知约束的双向生成对抗网络生成期望人物姿态图像的方法。采用给定人物的单个图像和期望姿态的二维骨架作为双向生成对抗网络的输入,生成具有该目标人物期望姿态的图像。将生成的期望姿态图像反映射回原始姿态图像,利用少量的图像以无监督学习方式进行学习,生成该人物期望姿态的高质量图像。提出的方法在DeepFashion公开数据集上进行了实验,结果表明,采用文中提出的方法生成的图像结构相似度(SSIM)比以往的方法提高了0.28,有效的提升了基于无监督学习的单人多姿态人物图像生成的质量。     

基于单片机的便携式电子血压计设计

基于单片机的便携式电子血压计设计,成功的设计出了4K字节FLASH储存器的低电压、高性能的CMOS8位微型处理器为控制核心的随身携带式电子血压计,本文将着重阐述电子血压计测量的原理以及软硬件设计方法。该血压计利用进口压力传感器将采集回的血管压力转化为电子信号,将模拟信号转成具体的数字信号后转回单片机,然后再经过计算、存储、显示等处理。这种家庭性的电子血压计的设计可以让人们使用起来更加方便,不耽搁高贵的时间。另外此系统设计中使用的被动组建和芯片有着投入成本低、功耗低、小型化等优点,这符合当今社会不断向前发展的前景。     

Research on algorithm for raster map multi-level sharing based on region incrementing color visual cryptography

In view of the characteristics of raster map,important data of the map were divided into different map layers by using the idea of map segmentation.A regional incremental color visual cryptographic scheme for raster map sharing was designed,and a model based on this scheme was proposed.At the same time,the specific application process was given.In order to solve the problem of large pixel expansibility and poor visual effect of image recovery in traditional regional incremental visual cryptography scheme,random numbers and the access structure optimization were used.Construction of the method was simple and the additional overhead of generating and saving the encryption matrix was avoided.The experimental results show that the scheme can achieve multilevel raster map sharing with the map’s perfect recovery,meanwhile the extension of the scheme pixel is greatly reduced.     

结合像素域JND阈值空间相关性与图像块类型的残差滤波算法

为了更好地将人眼视觉感知特性应用于视频压缩系统,从而去除视频的视觉冗余成分,在利用像素域的恰可察失真(Just Noticeable Distortion,JND)阈值的基础上,结合JND阈值的空间相关性与图像块类型,提出了一种改进的残差滤波算法。该算法首先在像素域计算出每个像素的JND阈值,然后在变换单元(Transform Unit,TU)中挖掘每个像素JND阈值与其周围邻近像素JND阈值之间的空间相关性,再利用索贝儿(Sobel)边缘检测算子将TU分成不同类型的图像块,并且计算出对应的复杂度因子,最后结合上述像素JND阈值的空间相关性和TU复杂度因子对TU残差进行滤波。提出的算法模型可以嵌入到高效率视频编码(High Efficiency Video Coding,HEVC)框架。实验结果表明,在全I帧配置下,提出的算法与标准算法HM16.0相比,在人眼主观感知质量基本一致的情况下,平均可节省16.1%的码率。