共查询到20条相似文献,搜索用时 15 毫秒
1.
A. M. Vil’kin I. V. Safonov M. A. Egorova 《Pattern Recognition and Image Analysis》2013,23(1):153-159
The ascending approach to segmentation of scanned documents in the area of background, text, and photographs is considered. In the first stage, the image is divided into blocks. For each block, a series of texture features is calculated. On the basis of these features, the type of the block is determined. Various positions and sizes of blocks, 26 texture features, and 4 algorithms of classification of blocks were considered. In the second stage, the type of block was corrected on the basis of the analysis of neighboring regions. For estimating the results, the error matrix and the ICDAR 2007 criterion are used. 相似文献
2.
针对雾天图像语义分割中分割精度不高的问题,基于空洞空间金字塔池化、Xception模块和残差网络,提出一种雾天图像语义分割算法.该算法一方面使用空洞空间金字塔池化和残差网络,以多个采样率的并行卷积以及卷积核大小为1×1的卷积对输入图像进行多尺度的上下文特征提取;另一方面,采用解码器结构使用预训练后的Xception模块对提取的特征进行分类,得到每一个像素的预测结果,用于增强分割边界的细化得到边界精细的分割结果.实验表明,所提算法在foggy cityscapes数据集上的平均交并比为73.03%、73.81%与74.50%,分割性能良好. 相似文献
3.
利用方向可控金字塔的方向可控性可以获取图像的方向信息,用来搜索最佳旋转变换参数,同时加入对平移量的搜索可以使这种方法能够处理具有平移和旋转变换的图像配准问题,采用互信息作为图像相似性度量的准则。实验表明该方法取得了较好的效果。 相似文献
4.
5.
为具有多种失真的图像提供一个无参考质量评价通用模型,提出了基于可控金字塔的评价算法.该算法结合自然场景图像的统计特性,利用可控金字塔变换对图像进行多尺度多方向的分解,并经过较小的训练,得到未失真图像子带系数分布特征模型.根据不同子带系数对图像降质的影响设置权值,计算量化后失真图像系数的实际分布与未失真自然场景子带系数分布特征模型的偏离程度,以度量图像的质量.实验结果表明,该算法能较好地符合人类视觉的主观评价. 相似文献
6.
Hossein Rabbani Author Vitae 《Pattern recognition》2009,42(9):2181-2193
This paper presents a new image denoising algorithm based on the modeling of coefficients in each subband of steerable pyramid employing a Laplacian probability density function (pdf) with local variance. This pdf is able to model the heavy-tailed nature of steerable pyramid coefficients and the empirically observed correlation between the coefficient amplitudes. Within this framework, we describe a novel method for image denoising based on designing both maximum a posteriori (MAP) and minimum mean squared error (MMSE) estimators, which relies on the zero-mean Laplacian random variables with high local correlation. Despite the simplicity of our spatially adaptive denoising method, both in its concern and implementation, our denoising results achieves better performance than several published methods such as Bayes least squared Gaussian scale mixture (BLS-GSM) technique that is a state-of-the-art denoising technique. 相似文献
7.
针对复杂视频场景中难以分割特定目标的问题,提出一种基于双重金字塔网络(DPN)的视频目标分割方法。首先,通过调制网络的单向传递让分割模型适应特定目标的外观。具体而言,从给定目标的视觉和空间信息中学习一种调制器,并通过调制器调节分割网络的中间层以适应特定目标的外观变化。然后,通过基于不同区域的上下文聚合的方法,在分割网络的最后一层中聚合全局上下文信息。最后,通过横向连接的自左而右结构,在所有尺度中构建高阶语义特征图。所提出的视频目标分割方法是一个可以端到端训练的分割网络。大量实验结果表明,所提方法在DAVIS2016数据集上的性能与较先进的使用在线微调的方法相比,可达到相竞争的结果,且在DAVIS2017数据集上性能较优。 相似文献
8.
针对现有语义分割算法参数量过多、内存占用巨大导致其很难满足自动驾驶需要等现实应用的问题,提出一种基于可分离金字塔模块(SPM)的新颖、有效且轻量的实时语义分割算法。首先,利用特征金字塔形式的分解卷积和扩张卷积来构建瓶颈结构,从而以一种简单但有效的方式提取局部和上下文信息;然后,提出基于计算机视觉注意力的上下文通道注意力(CCA)模块,来利用深层语义修改浅层特征图通道权重优化分割效果。实验结果显示:所提出的算法在Cityscapes测试集上以每秒91帧的速度达到了71.86%的平均交并比(mIoU)。相较高效残差分解卷积网络(ERFNet),所提算法mIoU提高了3.86个百分点,处理速度是其2.2倍;与最新的非局部高效实时算法(LRNNet)相比,所提算法mIoU略低0.34个百分点,但处理速度每秒上升了20帧。实验结果表明,所提算法有助于完成如自动驾驶中要求的高效、准确的街道场景图像分割任务。 相似文献
9.
10.
11.
针对RGB图像的实例分割任务在图像目标物体纹理相近但类别不同的区域可能出现分割错误的问题,引入Depth信息,结合RGB-D图像的三维几何结构特点,提出一种以双金字塔特征融合网络为框架的RGB-D实例分割方法.所提出的方法通过构建两种不同复杂度的金字塔深度卷积神经网络分别提取不同梯度分辨率大小的RGB特征及Depth特征,将对应分辨率大小的两种特征相加输入区域候选网络,以此改变输入区域候选网络层的共享特征,共享特征再经过分类、回归与掩码网络分支输出定位与分类结果,从而实现RGB-D图像的实例分割.实验结果表明,所提出的双金字塔特征融合网络模型能够完成RGB-D图像的实例分割任务,有效学习到深度图像与彩色图像之间的互补信息,与不包含Depth信息的Mask R-CNN相比,平均精度提高7.4%. 相似文献
12.
面向人机共融环境下机器智能对工具认知的需要,为提高家庭服务机器人的工具功能用途认知能力,设计一种基于深度几何特征空间金字塔池化的工具功用性建模与分类方法.离线训练阶段,考虑到各类工具在几何形态上的差异对工具自身更具表征性,首先,在各工具的深度图上提取多类几何特征,并融合形成工具特征图;然后,在工具特征图上提取多尺度特征块,并基于cciPCA的方法建立空间池化金字塔,从而构建最终的工具特征向量;最后,在高层语义空间上,利用SVM分类器训练工具分类识别模型.在线检测阶段,利用离线训练的工具分类模型对空间池化的样本进行分类测试.实验结果表明,所提方法能够实现家庭服务机器人对家庭日常工具的认知及分类识别,部分工具的识别精度可达97$%$及以上. 相似文献
13.
《Information Fusion》2002,3(3):203-214
Images acquired by different methodologies are fused to achieve more accurate information and a comprehensive understanding about the objects. With an increasing number of sensors involved, the signal acquisition and analysis system will benefit from the technique named sensor fusion or data fusion. Many successful applications have been found in the areas of computer vision, remote sensing, medical diagnosis, non-destructive testing, etc. Before the image fusion operation, a preprocessing, i.e. registration, needs to be performed, which ensures that all pixels are associated with corresponding physical points of the object. In this paper, we present an automatic two-dimensional image registration method by using a so-called steerable pyramid transform. Based on the steerability of the transform, image features along certain orientations can be obtained. Thereafter, an edge detector, such as Sobel, can further extract oriented edge features of the image. A coarse-to-fine procedure is adopted to utilize these features to achieve the registration result. The effectiveness of the proposed method is illustrated by the experimental results. 相似文献
14.
15.
16.
基于金字塔结构颜色特征的图像数据库检索 总被引:2,自引:1,他引:2
为了简单、准确定义图像的颜色特征,提出了基于区域块的颜色直方图技术。将目标集中在图像的中间区域,按照金字塔结构对图像分区并构建分区颜色直方图;相同区域内容在不同层中描述,从而可以在一定程度上解决普通直方图很难表达空间信息问题。 相似文献
17.
在苏木精-伊红(HE)染色病理图像中,细胞染色分布的不均匀和各类组织形态的多样性给病理图像的自动分割带来极大挑战。为解决该问题,提出了一种基于自监督学习的病理图像三步层次分割方法,对病理图像中各类组织进行由粗略到精细的全自动逐层分割。首先,根据互信息的计算结果在RGB色彩空间中进行特征选择;其次,采用K -means聚类将图像初步分割为各类组织结构的色彩稳定区域与模糊区域;然后,以色彩稳定区域为训练集采用朴素贝叶斯分类对模糊区域进行进一步分割,得到完整的细胞核、细胞质和胞外间隙这三类组织结构;最后,对细胞核部分进行结合形状和色彩强度的混合分水岭分割得到细胞核间的精确边界,进而量化计算细胞核个数、核占比、核质比等指标。对脑膜瘤HE染色病理图像的分割实验结果表明,所提方法对于染色和细胞形态差异保持较高的鲁棒性,各类组织区域分割误差在5%以内,在细胞核分割精度的对比实验中平均正确率在96%以上,满足临床自动图像分析的要求,其量化结果可以为定量病理分析提供依据。 相似文献
18.
Segmentation accuracy of dermoscopy images is important in the computer-aided diagnosis of skin cancer and a wide variety of segmentation methods for dermoscopy images have been developed. Considering that each method has its strengths and weaknesses, a novel adaptive segmentation framework based on multi-classification model is proposed for dermoscopy images. Firstly, five patterns of images are summarized according to the factors influencing segmentation. Then the matching relation is established between each image pattern and its optimal segmentationmethod. Next, the given image is classified into one of the five patterns by the multi-classification model based on BP neural network. Finally, the optimal segmentation method for this image is selected according to the matching relation, and then the image is effectively segmented. Experiments show that the proposed method delivers better accuracy and more robust segmentation results compared with the other seven state-of-the-art methods. 相似文献
19.
《Computer Vision, Graphics, and Image Processing》1990,49(3):346-356
Linked pyramid structures have proved to be a useful tool in digital image processing for many applicatins because of their ability to face problems at different levels of detail. Some variations suggested by usage to existent pyramid algorithms have been investigated for the segmentation of compact objects in noisy IR images. In particular, the efficacy of increasing the span at the very last iterations in order to correct the link deficiency of the boundary nodes is reported. We also report about a method which separate segment roots at any level in the pyramid and mergess the segments under the constraint of the maximum number of regions to be distinguished. The method is applied to IR image segmentation and comparative results are given. 相似文献