MPEG 4 video object segmentation grey consecutive region difference image.
标签MPEG -4视频对象分割灰度连续区域帧差图。
In this paper, a fast human face detection method based on difference image and multi-templates matching is proposed.
提出了一种用差影法与多模板匹配快速实现人脸检测的方法。
A human face detection method based on coarse segmentation of difference image and multi-template matching is proposed.
提出了基于差影法粗分割与多模板匹配的人脸检测方法。
A novel variable coefficient spatial filter is designed, which could effectively enhance the background difference image.
本文还设计了一种新的变系数空域滤波器,该滤波器能有效地对背景差图像进行显著增强。
A new filter algorithm named K neighborhood adaptive filter is proposed here for smoothing strong impulse disturbance in the difference image.
针对运动景物在差分图像中产生强脉冲干扰的特点,提出K邻域自适应滤波算法;
This method splits difference image into a set of subimages, and determines segmentation threshold of the whole scene by combining the thresholds of subimages.
该方法将差异图像分割成子图像,通过求子图像的局部阈值来确定差异图像的整体阈值。
The present invention aims to provide image processing device and method capable of generating a difference image at high speed without occurring misregistration.
本发明旨在提供能够以高速产生差图像而不会出现失配的图像处理设备和方法。
This system detected motion target using difference image method through threshold image segmentation, then automatically tracked target by using centroid tracking algorithm.
该系统运用差分图像法,通过阈值图像分割来检测运动目标,并采用形心跟踪算法来自动跟踪目标。
This algorithm can obtain the crossline position parameters using the inherent characters of crossline register marks and via the statistics difference image data or information.
该算法利用十字线图像的固有特征,通过统计差值图像信息,获得十字线的位置参数。
Tiny changes in a moving object can be detected by rebuilding the image background using the algorithm and calculating the ratio of pixel changes by means of the difference image algorithm.
用此算法重建背景图像以及用图像差分算法计算像素改变比例,能监测慢速、微量变化的运动物体。
The algorithm uses a difference image histogram of four directions to measure the influence of spatial correlation, and temporal correlation is measured by the histogram of frame difference.
该算法利用视频序列中各帧的四方向差分直方图度量隐写对空间相关性的影响,同时利用相邻帧帧间差分的直方图度量隐写对时间相关性的影响。
First based on the difference image and the skin color the face is detected and localized, then features are extracted by using improved SVD, last the features is classified by minimal distance.
首先基于差分图像和肤色信息检测出人脸,其次使用改进的奇异值分解方法提取面部特征,最后运用最小距离分类器进行识别。
Then, every frame is mapped from spatial domain to statistical domain incorporating the factors such as the difference image from the consecutive frames and the prior distribution of a pixel density.
然后,结合图像帧间的差分信息以及灰度分布的先验概率等因素将图像从空间域映射至统计域。
Then, every frame is mapped from spatial domain to statistical domain incorporating the factors such as the difference image from the consecutive frames and the prior distribution of a pixel density.
然后,结合图像帧间的差分信息以及灰度分布的先验概率等因素将图像从空间域映射至统计域。
应用推荐