Purpose: This study aimed to evaluate the ability of MRI-based intratumoral and peritumoral radiomics features of liver tumors to differentiate between hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC) and to predict ICC differentiation. Methods: This study retrospectively collected 87 HCC patients and 75 ICC patients who were confirmed pathologically. The standard region of interest (ROI) of the lesion drawn by the radiologist manually shrank inward and expanded outward to form multiple ROI extended regions. A three-step feature selection method was used to select important radiomics features and convolution features from extended regions. The predictive performance of several machine learning classifiers on dominant feature sets was compared. The extended region performance was assessed by area under the curve (AUC), specificity, sensitivity, F1 -score and accuracy. Results: The performance of the model is further improved by incorporating convolution features. Compared with the standard ROI, the extended region obtained better prediction performance, among which 6 mm extended region had the best prediction ability (Classification: AUC = 0.96, F1 -score = 0.94, Accuracy: 0.94; Grading: AUC = 0.94, F1 -score = 0.93, Accuracy = 0.89). Conclusion: Larger extended region and fusion features can improve tumor predictive performance and have potential value in tumor radiology.

Background: Primary liver tumors can be a serious threat to life and health. Early diagnosis may be life saving. Therefore, enhancing the accuracy of non-invasive early detection of liver tumors is imperative. Methods: Firstly, image enhancement was applied to augment the dataset, resulting in a total of 464 samples after employing seven data augmentation methods. Subsequently, the XGBoost model was utilized to construct and learn the mapping relationship between Computed Tomography (CT) and corresponding hyperspectral imaging (HSI) data. This model enables the prediction of HSI features corresponding to CT features, thereby enriching CT with more comprehensive hyperspectral information.Results: Four classifiers were employed to discern the presence of tumors in patients. The results demonstrated exceptional performance, with a classification accuracy exceeding 90%.Conclusions: This study proposes an artificial intelligence-based methodology that utilizes early CT radiomics features to predict HSI features. Subsequently, the results are utilized for non-invasive tumor prediction and early screening, thereby enhancing the accuracy of non-invasive liver tumor detection.

Diffusion tensor imaging (DTI) is a non-invasive magnetic resonance imaging technique and a special type of magnetic resonance imaging, which has been widely used to study the diffusion process in the brain. The signal-to-noise ratio of DTI data is relatively low, the shape and direction of the noisy tensor data are destroyed. This limits the development of DTI in clinical applications. In order to remove the Rician noise and preserve the diffusion tensor geometry of DTI, we propose a DTI denoising algorithm based on Riemann nonlocal similarity. Firstly, DTI tensor is mapped to the Riemannian manifold to preserve the structural properties of the tensor. The Riemann similarity measure is used to search for non-local similar blocks to form similar patch groups. Then the Gaussian mixture model is used to learn the prior distribution of patch groups. Finally, the noisy patch group is denoised by Bayesian inference and the denoised patch group is reconstructed to obtain the final denoised image. The denoising experiments of real and simulated DTI data are carried out to verify the feasibility and effectiveness of the proposed algorithm. The experimental results show that our algorithm not only effectively removes the Rician noise in the DTI image, but also preserves the nonlinear structure of the DTI image. Comparing to the existing denoising algorithms, our algorithm has better improvement of the principal diffusion direction, lower absolute error of fractional anisotropy and higher peak signal-to-noise ratio. © 2019, Springer-Verlag GmbH Germany, part of Springer Nature.

An optical single-channel color image encryption scheme based on chaotic fingerprint phase mask and diffractive imaging is proposed. In this proposed encryption scheme, the fingerprint used to generate the random phase masks is served as a secret key directly. Additionally, the random phase masks generated by the fingerprint, chaotic Lozi map, and secure hash algorithm (SHA-256) are used only as interim variables. With the help of the chaotic fingerprint phase masks placed at different diffraction distances, the color image that is encoded into a grayscale pattern by the phase-truncation technique is encrypted into a noise-like diffraction pattern. For decryption, the color image can be retrieved from the noise-like diffraction pattern by using an iterative phase retrieval algorithm, fingerprint, and phase keys generated from the encryption process. Since the fingerprint key shared by the sender and authorized receiver is strongly linked with the user and does not need to be transmitted over the open network, the security of this proposed encryption scheme can be greatly improved. Additionally, the parameters of the chaotic Lozi map and Fresnel diffraction distances can also provide additional security to the proposed encryption scheme. Furthermore, compared with the encryption schemes based on digital holography, the implementation of this proposed encryption scheme is relatively simple. The numerical simulations and analysis verify the feasibility, security, and robustness of this proposed encryption scheme. © 2023 Optica Publishing Group.

PURPOSE: To study the combined model of radiomic features and clinical features based on enhanced CT images for noninvasive evaluation of microsatellite instability (MSI) status in colorectal liver metastasis (CRLM) before surgery.; METHODS: The study included 104 patients retrospectively and collected CT images of patients. We adjusted the region of interest to increase the number of MSI-H images. Radiomic features were extracted from these CT images. The logistic models of simple clinical features, simple radiomic features, and radiomic features with clinical features were constructed from the original image data and the expanded data, respectively. The six models were evaluated in the validation set. A nomogram was made to conveniently show the probability of the patient having a high MSI (MSI-H).; RESULTS: The model including radiomic features and clinical features in the expanded data worked best in the validation group.; CONCLUSION: A logistic regression prediction model based on enhanced CT images combining clinical features and radiomic features after increasing the number of MSI-H images can effectively identify patients with CRLM with MSI-H and low-frequency microsatellite instability (MSI-L), and provide effective guidance for clinical immunotherapy of CRLM patients with unknown MSI status. © 2023. BioMed Central Ltd., part of Springer Nature.

Liver cancer is one of the leading causes of cancer death. Accurate and automatic liver tumor segmentation methods are urgent needs in clinical practice. Currently, Fully Convolutional Network and U-Net framework have achieved good results in medical image segmentation tasks, but there is still room for improvement. The traditional U-Net extracted a large number of low-level features, and the detailed features cannot be transmitted to deeper layers, resulting in poor segmentation ability. Therefore, this paper proposed a novel liver tumor segmentation network with contextual parallel attention and dilated convolution, called CPAD-Net. The pro-posed network applies a subsampled module, which has the same dimensionality reduction function as max -pooling without losing detailed features. CPAD-Net employs a contextual parallel attention module at skip connection. The module fuses contextual multi-scale features and extracts channel-spatial features in parallel. These features are concatenated with deep features to narrow the semantic gap and increase detailed informa-tion. Hybrid dilated convolution and double-dilated convolution are used in the encoding and decoding stages to expand the network receptive field. Dropout is added after each hybrid dilated convolution block to prevent overfitting. The efficacy of the proposed network is proved by widespread experimentation on two public datasets (LiTS2017 and 3Dircadb-01) and a clinical dataset from the Affiliated Hospital of Hebei University. The proposed network achieved Dice scores of 74.2%, 73.7% and 73.26%. The experimental results show that the proposed network outperforms most segmentation networks.

单样本人脸识别是人脸识别在实际应用中面临的挑战性问题之一，虽然深度学习在人脸识别方面取得突破性进展但其性能依赖海量标注性数据，故其在单样本上性能有限。而传统浅层特征对有标注的数据量需求不高，但因单样本数据缺少类内变化其性能有限，提出一种改进加权投票的PCA-Net多特征融合算法。在数据集方面，利用LU分解生成虚拟样本扩展数据集；根据PCA-Net特征下样本的相关性细化数据集，实现对数据集初步特征提取和筛选；在细化数据集上提取多LBP特征并与PCA-Net特征进行加权投票。在AR、Extended Yale B、CMU-PIE三个数据库上的实验结果表明，所提方法提高了单样本人脸识别性能。

合成孔径雷达（Synthetic Aperture Radar,SAR）遥感与光学遥感相比,能够全天时、全天候、全地域进行实时数据采集,成像效果不受时间、气候和地域的影响。同时,合成孔径雷达成像分辨率与目标距离无关,在国防和民生方面具有非常重要的作用。由于其相干成像机制,合成孔径雷达图像中不可避免地会产生斑点噪声。斑点噪声是一种依赖于观测信号的颗粒状噪声,是相干成像系统中固有的噪声,相干斑的存在降低了图像质量,影响了后续的分割与解译,因此,研究适合合成孔径雷达图像特点的去噪算法尤为重要。低秩重构及成分分析理论近年来备受国内外学者关注,具有重要的理论价值和应用潜力,但用于SAR图像去噪仍存在许多问题需要解决。本文在前人的基础上研究基于低秩重构及成分分析理论的SAR图像斑点噪声降低方法,本文的主要创新性研究成果如下:（1）提出了一种基于纹理强度和加权核范数最小化的SAR图像盲去噪算法。首先分析了图像的结构特征,利用梯度协方差矩阵的迹对图像结构进行表征,从而得到表征图像纹理结构的纹理强度参考量。然后利用纹理强度选择低秩子块,由选择的低秩子块估计图像的噪声水平。利用估计的噪声方差以及观测噪声图像子块的奇异值,可以估计未知的原始无噪声图像的奇异值。考虑到大的奇异值代表原始数据的主要成分,应尽量保留,而小的奇异值代表原始数据的次要成分,应尽量收缩。所以,对不同奇异值赋予不同的权重收缩系数,因而权重收缩系数应采用非递减的顺序排列。加权核范数最小化问题本是非凸优化问题,但是当权重系数按照非递减顺序排列时,该问题存在最优的解析解。因此加权核范数最小化可以用于图像降噪并得到最优解。实验结果表明,基于加权核范数最小化的盲去噪方法用于处理SAR图像时在主观和客观方面都具有较强的竞争力。更重要的是,采集的SAR图像噪声污染水平通常是未知的,我们提出一种针对SAR图像的盲去噪框架。一方面可以利用低秩子块估计噪声方差,另一方面,可以利用选择的去噪算法调整噪声方差从而使得去噪算法达到最优的去噪效果。基于纹理强度和加权核范数最小化的SAR图像盲去噪方法具有更高的鲁棒性,在工程实践中具有很好的应用前景。（2）提出了一种基于加权低秩恢复的SAR图像去噪算法。基于SAR图像中存在结构冗余,利用低秩恢复方法恢复原始数据。对每一个像素利用排序绝对差值计算其污染概率,将污染概率作为约束条件加入低秩表示模型,从而约束恢复图像的正则性。利用加权平均进一步抑制噪声并采用改进算法来减少局部滤波和全局恢复之间的差异,得到去噪后的图像。实验结果表明,该算法在仿真斑点噪声模型中得到了较高的客观指标,尤其是在背景对比度方面有显著提高,这说明对SAR图像的辐射保真度保护较好。从仿真图像和真实图像的局部放大图像看出,与同类方法相比,基于加权低秩恢复的SAR图像去噪方法在有效平滑噪声的同时,具有更好的纹理保护能力。（3）提出了一种分组主成分分析和导向滤波的SAR图像去噪算法。将待处理像素与其邻域像素组成一个向量,在局部窗口选择与该向量相似的向量组成训练集。用块匹配测量作为相似性判定准则,寻找与待处理向量空间结构相似的向量。对训练集做主成分分析（Principal Component Analysis,PCA）,然后对变换系数进行收缩。利用局部分组,能够准确计算变量的局部信息,从而在PCA域收缩系数后能更好地保护图像的边界结构。由于强噪声的存在会造成局部向量分组错误以及PCA变换矩阵计算偏差,因此仍有噪声残留。导向滤波不仅具有良好的边缘保持平滑特性,而且可以高效计算,因此采用导向滤波进一步去噪。实验结果表明,该方法在保护边界方面优于基于非局部的方法,同时避免产生伪影。（4）提出了一种将小波-轮廓波变换与迭代循环移位相结合的SAR图像去噪算法。考虑到目前存在的多尺度几何变换方法往往缺乏移不变性并且冗余度较高,轮廓波变换不具备移不变性,直接对变换后的系数设定阈值时会产生伪吉布斯现象,因此采用迭代循环移位代替多次移位取平均值,从而减少伪吉布斯现象。对实际的SAR图像进行去噪,实验表明,小波-轮廓波加上迭代循环移位方法不仅在客观性能参数方面具有较大的改进,而且去噪之后的图像视觉效果优于其他同类方法,可降低伪影信息。

图像融合技术是利用同一场景或同一目标的多幅输入图像之间信息的互补性,将其整合为一幅图像的过程,该技术能够提高图像中场景或目标的清晰度,并为后续的目标识别、分类等工作提供了更精确可靠的图像支持。目前,该技术已成功应用于军事、农业、安全和监控等众多领域。本文主要研究基于形态分量分析（Morphological Component Analysis,MCA）的多聚焦图像和医学图像融合算法,具体内容如下:第一,改进了形态分量分析算法中的迭代次数和过完备字典的选取方法。针对MCA的迭代次数过小会导致图像分离结果不准确,过大会降低算法效率的问题,本文提出利用小波变换和中值绝对偏差估计源图像的噪声值,并将其作为图像分解过程中残差的阈值,以确保该算法能够根据图像自身特性自适应地选择分解迭代次数。另外,针对MCA采用的固定字典自适应能力较弱的问题,本文提出采用K-SVD（K-Singular Value Decomposition）算法训练字典,以更好地利用卡通分量和纹理分量的特性。实验结果表明,改进MCA算法获得的融合图像在主观视觉和客观评价指标中均有更优的表现。第二,提出基于改进MCA和非下采样剪切波变换（Non-subsampled Shearlet Transform,NSST）的多聚焦图像融合算法。为弥补MCA细节表现能力不足的缺陷,本文提出采用NSST分解卡通分量,以捕捉其各个尺度各个方向的细节特征。另外,提出空间频率和脉冲耦合神经网络（Pulse Coupled Neural Network,PCNN）相结合的纹理融合规则,并结合人类视觉特性提出改进的拉普拉斯能量和（Sum-Modified-Laplacian,SML）加权平均的低频融合规则。实验结果表明,该算法能够准确判断多聚焦图像的清晰区域,并减少了传统变换域算法中出现的伪影、扭曲等影响图像质量的现象。第三,提出基于改进MCA和多目标粒子群优化（Multi-Objective Particle Swarm Optimization,MOPSO）的医学图像融合算法,主要利用MOPSO优化卡通融合规则及图像重建规则中的加权系数,并结合IHS（Intensity,Hue,Saturation）颜色模型将该算法应用到彩色医学图像融合中。实验结果证明,该算法在灰度医学图像融合及彩色医学图像融合中均能够得到优于对比算法的融合结果。