【摘要】 给出了这类图像的一种图像模型,并在此基础上,提出了一种新的分割和检测奇异点的方案。该方案对原图像进行非分除小波变换,以得到带通图像,使得带通图像上的奇异点得到增强,同时背景和噪声得到抑制。将所得到的带通图像分割成互相重叠的方形区域,通过计算每个区域的扭曲度(Skerness)和峭度(Kurtosis)特征来判断该区域分布的非对称性和拖尾程度,并将具有较高值的区域标记为感兴趣区域ROI(Regions Of Interest)。在ROI中,如果能量特征超过某一给定阀值,则被视为准奇异点,形成二值图像。对二值图像进行数学形态处理的检测结果进一步证明了该奇异点分割和检测方案的有效性。更多还原

【Abstract】 In natural images, the small man-made objects can be viewed as anomalies on the background. In medical images, so the same as the tiny calcium deposits in tissues. An image model for the conditions above is given, based on which, a new scheme for segmenting and detecting anomalies is proposed. The image is first processed using undecimated wavelet transform to get the band pass image in which the anomalies are enhanced and the background is suppressed. Then, the band pass image is divided into overlapping square regions in which skew ness and kurtosis as measures of the asymmetry and impulsiveness of the distribution are estimated. Regions with high positive values are marked as regions of interest (ROI). In ROI, pixels with energy greater than a certain threshold are candidates of the anomalies, which are on in a binary image. Finally, binary morphological operations are applied to the binary image to get the detection result. The experiment results show that the new scheme is an effective means for anomalies detection.更多还原