Author Affiliations
Abstract
School of Astronautics, Harbin Institute of Technology, Harbin, Heilongjiang 150000, P. R. China
Photoacoustic imaging (PAI) is a noninvasive emerging imaging method based on the photoacoustic effect, which provides necessary assistance for medical diagnosis. It has the characteristics of large imaging depth and high contrast. However, limited by the equipment cost and reconstruction time requirements, the existing PAI systems distributed with annular array transducers are difficult to take into account both the image quality and the imaging speed. In this paper, a triple-path feature transform network (TFT-Net) for ring-array photoacoustic tomography is proposed to enhance the imaging quality from limited-view and sparse measurement data. Specifically, the network combines the raw photoacoustic pressure signals and conventional linear reconstruction images as input data, and takes the photoacoustic physical model as a prior information to guide the reconstruction process. In addition, to enhance the ability of extracting signal features, the residual block and squeeze and excitation block are introduced into the TFT-Net. For further efficient reconstruction, the final output of photoacoustic signals uses ‘filter-then-upsample’ operation with a pixel-shuffle multiplexer and a max out module. Experiment results on simulated and in-vivo data demonstrate that the constructed TFT-Net can restore the target boundary clearly, reduce background noise, and realize fast and high-quality photoacoustic image reconstruction of limited view with sparse sampling.
Deep learning feature transformation image reconstruction limited-view measurement photoacoustic tomography Journal of Innovative Optical Health Sciences
2024, 17(3): 2350028
Author Affiliations
Abstract
1 State Key Laboratory of Extreme Photonics and Instrumentation, College of Optical Science and Engineering, Zhejiang University, Hangzhou 310027, P. R. China
2 ZJU-Hangzhou Global Scientific and Technological Innovation Center, Hangzhou 311200, P. R. China
3 Collaborative Innovation Center of Extreme Optics, Shanxi University, Taiyuan, Shanxi 030006, P. R. China
4 Advanced Biomedical Imaging Facility-Wuhan National Laboratory for Optoelectronics, Huazhong University of Science and Technology, Wuhan, Hubei 430074, P. R. China
Structured illumination microscopy (SIM) achieves super-resolution (SR) by modulating the high-frequency information of the sample into the passband of the optical system and subsequent image reconstruction. The traditional Wiener-filtering-based reconstruction algorithm operates in the Fourier domain, it requires prior knowledge of the sinusoidal illumination patterns which makes the time-consuming procedure of parameter estimation to raw datasets necessary, besides, the parameter estimation is sensitive to noise or aberration-induced pattern distortion which leads to reconstruction artifacts. Here, we propose a spatial-domain image reconstruction method that does not require parameter estimation but calculates patterns from raw datasets, and a reconstructed image can be obtained just by calculating the spatial covariance of differential calculated patterns and differential filtered datasets (the notch filtering operation is performed to the raw datasets for attenuating and compensating the optical transfer function (OTF)). Experiments on reconstructing raw datasets including nonbiological, biological, and simulated samples demonstrate that our method has SR capability, high reconstruction speed, and high robustness to aberration and noise.
Structured illumination microscopy image reconstruction spatial domain digital micromirror device (DMD) Journal of Innovative Optical Health Sciences
2024, 17(2): 2350021
1 重庆大学ICT研究中心光电技术及系统教育部重点实验室,重庆 400044
2 重庆大学工业CT无损检测教育部工程研究中心,重庆 400044
针对相对平行直线扫描CT(PTCT)图像重建存在的有限角伪影问题,提出一种学习局部和非局部正则项的深度迭代展开方法。该方法将具有固定迭代次数的梯度下降算法迭代展开到神经网络,利用具有坐标注意力(CA)机制的卷积模块和Swin-Transformer模块作为迭代模块交替级联部署,构成端到端的深度重建网络。卷积模块学习局部正则化,其中CA用于减少图像过平滑;Swin-Transformer模块学习非局部正则化,提高网络对图像细节的恢复能力;在相邻模块间,使用迭代连接(IC)增强模型提取深层特征的能力,提高每次迭代的效率。通过消融实验验证了网络各部分的有效性,并在两种类型的数据集上进行实验,结果证明了本文方法的效果。实验结果表明,本文方法在抑制PTCT重建图像有限角伪影的同时,能较好地保留重建图像细节,提高重建图像质量。
X射线光学 计算机断层成像 相对平行直线扫描 图像重建 有限角 深度学习
Author Affiliations
Abstract
1 School of Optoelectronic Science and Engineering, Soochow University, Suzhou 215006, China
2 Key Laboratory of Modern Optical Technologies of the Ministry of Education, Soochow University, Suzhou 215006, China
The source’s energy fluctuation has a great effect on the quality of single-pixel imaging (SPI). When the method of complementary detection is introduced into an SPI camera system and the echo signal is corrected with the summation of the light intensities recorded by two complementary detectors, we demonstrate, by both experiments and simulations, that complementary single-pixel imaging (CSPI) is robust to the source’s energy fluctuation. The superiority of the CSPI structure is also discussed in comparison with previous SPI via signal monitoring.
computational imaging image reconstruction complementary detection correlation function Chinese Optics Letters
2024, 22(3): 031101
1 西南交通大学轨道交通运载系统全国重点实验室摩擦学研究所,四川 成都 610031
2 西南交通大学唐山研究院,河北 唐山 063000
3 中国工程物理研究院核物理与化学研究所,四川 绵阳 621900
4 北京航空航天大学机械工程及自动化学院,北京 100191
中子偏置CT(computed tomography)扫描是一种有效的大尺寸样品层析检测方法,但投影数据截断会导致较大的CT系统转台旋转中心标定误差,严重影响成像质量。基于投影数据对称性原理,提出了一种计算旋转中心左侧和右侧投影数据和之间方差的偏置CT扫描旋转中心精确标定算法。设计了对称补数据重建算法和投影数据预处理重建算法,验证得到,对称补数据重建算法对旋转中心标定误差更为敏感,较小的误差值会导致补齐后投影数据出现拼接缝以及拼接错位问题。提出了一种中子投影数据噪声仿真方法,设计的三维仿真模体验证了所提标定算法与投影数据预处理重建算法在不同旋转中心偏置大小以及不同强度投影噪声条件下的性能优势。基于反应堆中子源开展了中子偏置CT扫描成像验证实验,获得了样品清晰的内外部结构细节,中子CT成像系统的成像视野扩大了31.4%。
计算机断层扫描成像 图像重建技术 中子 偏置扫描 旋转中心
雷腾 1,2,3张义民 1,2,3马一哲 1,2,3丁学专 1,2,3[ ... ]王世勇 1,2,3,*
1 中国科学院 上海技术物理研究所,上海200083
2 中国科学院大学,北京100049
3 中国科学院 红外探测与成像技术实验室,上海20008
低采样率下的高质量鬼成像(GI)对于科学研究和实际应用具有重要意义,为了在低采样率条件下重建高质量图像,提出了一种高质量的被动式压缩鬼成像重构算法(PCGI-LRC)。基于图像的非局域相似块堆叠而成的矩阵具有低秩和稀疏奇异值的假设,从理论和实验上证明了一种对最小二乘问题与非局域相似块低秩近似问题进行联合迭代求解的方法,能够在低采样率(6.25%~50%)条件下实现高质量鬼成像。实验结果表明:与基于稀疏基约束的GI(GI-SBC)和基于全变分约束的GI(GI-TVC)相比,PCGI-LRC在峰值信噪比、结构相似性系数和视觉观测等方面均更优,在抑制重构噪声的同时保持了目标的细节信息,其中PSNR提升效果优于1.1 dB,SSIM提升效果优于0.04。
鬼成像 图像重构 图像压缩 单像素成像 ghost imaging image reconstruction image suppression single pixel imaging
图像压缩感知是一种在欠采样条件下尽可能重构原始图像的技术。为解决大部分基于卷积神经网络(CNN)框架的图像压缩感知方法容易受到卷积感受野的限制、对全局信息的关注较少的问题, 提出了基于Swin Transformer的图像压缩感知重构网络。网络使用卷积层对图像进行采样, 然后使用自注意力机制和残差结构结合的残差Swin Transformer组(RSTG)结构来关注图像的细节。实验结果表明, 基于Swin Transformer的图像压缩感知重构网络可以充分利用图像的先验信息, 进一步提高图像压缩感知的重构精度, 并获得比其他压缩感知方法更好的重构性能和视觉效果。
压缩感知 图像重构 自注意力机制 残差 compressive sensing image reconstruction self-attention mechanism residual
1 中科技术物理苏州研究院,江苏 苏州215000云遥动力科技(苏州)有限公司,江苏 苏州215000
2 中科技术物理苏州研究院,江苏 苏州215000中国科学院上海技术物理研究所,上海200083云遥动力科技(苏州)有限公司,江苏 苏州215000
风云四号A卫星(FY-4A)部分通道的图像分辨率较低,影响风云卫星多通道协同监测能力。针对该问题,提出一种改进的超分辨率生成对抗网络(Enhanced Super-Resolution Generative Adversarial Networks, ESRGAN)算法来实现FY-4A影像的超分辨率重建。该方法基于ESRGAN的生成器架构,利用迁移学习策略,将预训练权重作为模型的初始值,设计了一组包含5层空洞卷积层的残差密集模块(Residual-in-Residual Dense Block, RRDB),并优化了损失函数。结果表明,在4倍影像分辨率重建下,相比于ESRGAN算法,改进的ESRGAN算法在峰值信噪比(Peak Signal-to-Noise Ratio, PSNR)、结构相似性(Structural Similarity,SSIM)、空间相关系数(Correlation Coefficient,CC)上分别提高了0704、0029、0002,均方根误差(Root-Mean-Square Error,RMSE)降低了10%,且重建影像更加清晰自然,纹理更加细致。
风云四号 深度学习 图像重建 超分辨率 Fengyun-4 deep learning image reconstruction super resolution
光子学报
2023, 52(11): 1110004