基于改进YOLOv5的荧光图像细胞智能检测研究

为解决人工对荧光原位杂交(Fluorescence In Situ Hybridization, FISH)荧光图像进行结果判读存在的效率低、劳动强度大等问题,针对FISH荧光图像细胞智能检测提出一种融合空域图像增强的改进YOLOv5算法。算法在原始YOLOv5神经网络模型基础上,加入了空域图像增强模块,并选择了模块最佳增强系数,扩大了模型对荧光图像的对比度适应范围,提高了模型的特征提取能力和细胞检测准确率。实验结果显示,改进YOLOv5模型的平均精度均值(Mean Average Precision, mAP)为0.983,达到了比原始模型更优的训练效果和收敛速度,并且,改进YOLOv5模型的细胞识别率达到91.65%,比原始YOLOv5模型提升了9.19%。将细胞智能检测算法嵌入自主开发的荧光图像智能检测软件,结合荧光点检测算法,可给出有效判读结果。     

Ultrasensitive Telomerase Activity Detection in Circulating Tumor Cells Based on DNA Metallization and Sharp Solid‐State Electrochemical Techniques

Being considered a “liquid biopsy”, circulating tumor cell (CTC) quantification is of great interest for evaluating cancer dissemination, predicting patient prognosis, and also for the evaluation of therapeutic treatments, representing a reliable potential alternative to invasive biopsies and subsequent proteomic and functional genetic analysis. Compared to a biopsy, the gold standard of current cancer diagnosis, an important characteristic of a blood test is that it is safe and can be performed at many points during the disease, allowing the development of appropriate therapy modifications and potentially improving patient's quality of life. In this work, an ultrasensitive electrochemical telomerase activity‐sensing strategy is presented that utilizes DNA‐templated deposition of silver nanoparticles as electroactive labels through a highly sharp solid‐state Ag/AgCl reaction with DNA exonuclease III‐assisted background current suppression. This nanoparticle‐mediated signal amplification resulted in significantly decreased detection limit, which is better than the vast majority of reported methods and achieves a sensitivity comparable to the conventional telomeric repeat amplification protocol (TRAP). This work may pave a new PCR‐free way for the detection of telomerase activity in CTCs via a noninvasive routine blood test for point‐of‐care diagnosis and individualized treatment of cancer.     

Quantitative Comparison of Spot Detection Methods in Fluorescence Microscopy

Quantitative analysis of biological image data generally involves the detection of many subresolution spots. Especially in live cell imaging, for which fluorescence microscopy is often used, the signal-to-noise ratio (SNR) can be extremely low, making automated spot detection a very challenging task. In the past, many methods have been proposed to perform this task, but a thorough quantitative evaluation and comparison of these methods is lacking in the literature. In this paper, we evaluate the performance of the most frequently used detection methods for this purpose. These include seven unsupervised and two supervised methods. We perform experiments on synthetic images of three different types, for which the ground truth was available, as well as on real image data sets acquired for two different biological studies, for which we obtained expert manual annotations to compare with. The results from both types of experiments suggest that for very low SNRs $(approx 2)$, the supervised (machine learning) methods perform best overall. Of the unsupervised methods, the detectors based on the so-called $h$ -dome transform from mathematical morphology or the multiscale variance-stabilizing transform perform comparably, and have the advantage that they do not require a cumbersome learning stage. At high SNRs $(> 5)$, the difference in performance of all considered detectors becomes negligible.      

多分类器联合虚警可控的海上小目标检测方法

模式识别技术已经广泛应用于海上目标检测,其中二分类的模式识别算法在处理该问题时会面临类别非均衡的困境。传统方法一般通过添加人工仿真目标回波扩充目标数据集,检测结果容易受到仿真精度的影响,且增加算法的复杂度。该文提出一种基于多分类思想的多特征海上小目标智能检测方法,先对海杂波数据与目标数据进行多维特征提取,构建高维特征空间;再基于多分类思想中的“1对1”方法,将海杂波特征空间划分成多个子空间,每个杂波子空间与目标数据特征空间等大,构造多个二分类器进行联合判决。该文选取的二分类器为改进的双参数K近邻 (K-NN)算法,可有效调节虚警率。经冰多参数成像X波段雷达(IPIX)数据集验证,所提方法在观测时间为1.024 s时获得了82.40%的检测概率,与基于K-NN的检测器做比较,获得了2%的性能提升。     

基于结构照明的超分辨荧光显微成像重建算法

由于具有低光毒性、高速宽视场以及多通道三维超分辨成像能力,超分辨结构照明显微术(SR-SIM)特别适合用于活细胞中动态精细结构的实时检测研究。超分辨结构照明显微图像重建算法(SIM-RA)对SR-SIM的成像质量具有决定性影响。本文首先简要介绍了超分辨显微术的发展现状,阐述了研究SR-SIM图像重建算法的必要性;然后介绍了SR-SIM的成像原理,并重点介绍了SR-SIM图像重建算法,包括SR-SIM中频繁使用的去卷积重建算法、SR-SIM校准与重建过程中参数值获取的算法,以及目前发展的超分辨结构照明显微图像重建算法,并介绍了SR-SIM工具箱;最后总结了当前发展超分辨结构照明显微图像重建算法需解决的5个问题。     

Hyperspectral Fluorescence Imaging for Mouse Skin Tumor Detection

This paper presents a hyperspectral imaging technique based on laser‐induced fluorescence for non‐invasive detection of tumorous tissue on mouse skin. Hyperspectral imaging sensors collect image data in a number of narrow, adjacent spectral bands. Such high‐resolution measurement of spectral information reveals contiguous emission spectra at each image pixel useful for the characterization of constituent materials. The hyperspectral image data used in this study are fluorescence images of mouse skin consisting of 21 spectral bands in the visible spectrum of the wavelengths ranging from 440 nm to 640 nm. Fluorescence signal is measured with the use of laser excitation at 337 nm. An acousto‐optic tunable filter (AOTF) is used to capture images at 10 nm intervals. All spectral band images are spatially registered with the reference band image at 490 nm to obtain exact pixel correspondences by compensating the spatial offsets caused by the refraction differences in AOTF at different wavelengths during the image capture procedure. The unique fluorescence spectral signatures demonstrate a good separation to differentiate malignant tumors from normal tissues for rapid detection of skin cancers without biopsy.     

荧光寿命成像显微的频域零差法测量及数据处理

介绍了一种实现荧光寿命成像显微技术（FLIM）的频域零差法，并改进了测量数据的逐点分析法，设计了相应的数据处理软件．通过对模拟数据的处理表明，完全满足FLIM数据处理的要求。     

荧光寿命成像显微的频域外差法测量及数据处理方法研究

介绍了一种实现荧光寿命成像显微的方法──频域外差法，研究并改进了相位调制度数据的逐点分析法，设计了相应的数据处理软件。通过对模拟数据的处理表明，该分析法具有很高的处理精度。     

一种基于MRI图像的脑肿瘤组合分割法

从模糊的MRI脑肿瘤的图像中准确找出肿瘤,供医学临床使用。分别采用了2种分割方法：一种是改进的阈值方法,即在进行最大方差阈值法之前首先采用手动阈值法;另一种是采用圆形结构元素作为种子的形态学分割方法。这2种方法均实现了从低对比度MRI脑图像提取目标的要求。灵活使用阈值法和形态学分割方法与医学图像中,具有一定的临床实用价值。     

Intracranial Epileptic Seizures Detection Based on an Optimized Neural Network Classifier

Automatic identification of intracranial electroencephalogram (iEEG) signals has become more and more important in the field of medical diagnostics. In this paper, an optimized neural network classifier is proposed based on an improved feature extraction method for the identification of iEEG epileptic seizures. Four kinds of entropy, Sample entropy, Approximate entropy, Shannon entropy, Log energy entropy are extracted from the database as the feature vectors of Neural network (NN) during the identification process. Four kinds of classification tasks, namely Pre-ictal v Post-ictal (CD), Pre-ictal v Epileptic (CE), Post-ictal v Epileptic (DE), Pre-ictal v Post-ictal v Epileptic (CDE), are used to test the effect of our classification method. The experimental results show that our algorithm achieves higher performance in all tasks than previous algorithms. The effect of hidden layer nodes number is investigated by a constructive approach named growth method. We obtain the optimized number ranges of hidden layer nodes for the binary classification problems CD, CE, DE, and the multitask classification problem CDE, respectively.     

基于改进RetinaNet的SAR图像目标检测方法

近年来,深度学习方法在合成孔径雷达(SAR)图像目标检测中得到了广泛的应用.船舶出现在近海、港口、岛礁、远洋等各种场景中,同时海洋环境复杂多变,使得船舶目标检测很难排除混乱背景的干扰.对于大纵横比、任意方向、密集分布的目标,精确定位变得更加复杂.本文基于深度学习的方法提出用于SAR图像目标检测的改进RetinaNet模...     

基于集成 PU 学习数据流分类的入侵检测方法

入侵检测问题可以模型化为数据流分类问题,传统的数据流分类算法需要标注大量的训练样本,代价昂贵,降低了相关算法的实用性。在PU学习算法中,仅需标注部分正例样本就可以构造分类器。对此本文提出一种动态的集成PU学习数据流分类的入侵检测方法,只需要人工标注少量的正例样本,就可以构造数据流分类器。在人工数据集和真实数据集上的实验表明,该方法具有较好的分类性能,在处理偏斜数据流上优于三种PU 学习分类方法,并具有较高的入侵检测率。     

基于Harris角点的彩色图像文字检测

提出了一种基于Harris角点的文字检测算法.首先,根据彩色图像和视频中文字区域和背景之间的颜色分量大小的对比,利用Harris角点检测算法,得到图像的角点分布图;然后对图像进行滤波,去除相对孤立的角点;将角点图像进行二值化,利用形态学运算将角点聚合形成区域;对区域进行轮廓跟踪算法,得到文字区域的初定位图像;最后对文字区域进行验证,得到最终结果.实验证明该算法具有较高的准确性.     

基于迁移学习的SAR图像目标检测

针对深度卷积神经网络训练需要大数量样本,采用迁移学习的方法辅助网络训练,解决了SAR图像样本不足的问题。通过控制对比实验,对每个卷积块权重进行迁移与分析,使用微调与冻结相结合的训练方式有效提高网络的泛化性与稳定性;然后根据目标检测任务的时效性对网络模型进行改进,提高了网络检测速度的同时减少了网络参数;最后结合复杂场景杂波切片对网络进行训练,降低了背景杂波的虚警目标数量,复杂多目标场景的检测结果表明所提出方法具有较好的检测性能。     

基于YOLOX网络的SAR图像舰船目标检测

针对SAR图像舰船目标尺寸大小不一、舰船分布密集、背景复杂等问题,本文提出一种改进YOLOX网络并用于SAR图像舰船目标检测。该网络包括主干特征提取网络、加强特征提取网络、解耦头、预测框优化及损失计算等4个部分。与常规YOLOX网络相比,本文作了如下改进：首先,在主干特征提取网络中,3个基础特征层之后都添加了CA模块;在加强特征提取网络中,两处下采样之后也都添加了CA模块。以强化对SAR图像中重要区域的特征提取。其次,在框回归损失函数中,引入CIOU替代IOU,以更好地利用预测框和真实框之间的相对位置信息和形状信息,提升预测框回归精度。本文基于AIR-SARSHIP-2.0数据集进行了大量的舰船目标检测实验,并选择了Faster-RCNN、YOLOv3和常规YOLOX等3种网络与本文的改进YOLOX网络进行对比。实验结果表明,本文的改进YOLOX网络整体性能优于其他3种对比网络,有更少的虚警和漏警、更高的检测精度。     

基于多通道特征的SAR图像变化检测

为充分利用SAR图像的细节信息,提高SAR图像变化检测的检测精度及抗噪性能,提出一种基于多通道特征的SAR图像变化检测方法。该方法提出了一种适用于SAR图像的变化检测一体化框架,首先,为了在抑制相干斑噪声的同时尽可能多地保留SAR图像的边缘及局部信息,引入引导图像滤波方法;其次,提取8个通道特征,充分利用了图像的细节信息,获得了性能良好的差异图;最后,利用主成分分析(PCA)和K-means聚类进行差异图分析,得到最终的变化信息。实验结果表明,该方法有效提高了检测精度,并且具有良好的抗噪性能。     

DNA‐Templated Magnetic Nanoparticle‐Quantum Dot Polymers for Ultrasensitive Capture and Detection of Circulating Tumor Cells

Circulating tumor cell (CTC) enumeration and analysis has emerged as an important platform for cancer diagnosis and prognosis. A great challenge, however, is to efficiently capture low abundant CTCs with high purity from blood samples in a rapid and high‐throughput manner for accurate and sensitive CTC detection. Herein, a new class of DNA‐templated magnetic nanoparticle‐quantum dot (QD)‐aptamer copolymers (MQAPs) is developed for rapid magnetic isolation of CTCs from human blood with high capture efficiency and purity approaching 80%. The phenotype of CTCs is simultaneously profiled with QD photoluminescence (PL) at single cell level. These MQAPs are constructed through hybridization chain reaction to achieve amplified magnetic response, extraordinary binding selectivity for target cells over background cells, and ultra bright ensemble QD PL for single cell detection. MQAPs are free from nonspecific binding that would otherwise compromise the capture purity of target cells. As a result, facile isolation and enumeration of rare CTCs in blood samples could be achieved in 20 min with high sensitivity and accuracy.     

Targeting of CXCR1 on Osteosarcoma Circulating Tumor Cells and Precise Treatment via Cisplatin Nanodelivery

Metastasis and chemotherapy resistance are the key factors affecting the effectiveness of osteosarcoma (OS) treatments. CXCR1 overexpression is found to be closely related to chemotherapy resistance and anoikis resistance in OS cell subtypes with high metastasis potential. Further study demonstrates that CXCR1 is highly expressed on circulating tumor cell (CTC)‐derived cells with cancer stem cell characteristics. Then, a CXCR1 targeting peptide is designed and synthesized to competitively inhibit the IL‐8/CXCR1 pathway and to improve the cisplatin sensitivity of CTCs. Fluorescence‐labeled magnetic nanoparticles (NPs) with pH‐responsive cisplatin release are fabricated and linked with the CXCR1 targeting peptide (Cis@MFPPC). Results demonstrate that CTC survival could be inhibited effectively by the targeting nanoparticles in vivo. Cis@MFPPC can also inhibit OS growth and pulmonary metastasis in an orthotopic model and patient‐derived tumor xenograft model. This study verifies the clinical significance of CXCR1 as a therapeutic target and provides a drug delivery NP system for precise treatment of OS.     

基于红外图像的避雷器故障检测方法

在电力系统中,利用计算机视觉和图像处理技术对避雷器进行故障检测,在保障电力系统的安全运行方面具有非常重要的作用。提出了一种基于红外图像的避雷器故障检测方法。该方法首先对输入图像进行预处理,利用尺度不变特征变换（Scale-invariant feature transform,SIFT）描述子和K-means++算法训练视觉字典精确定位避雷器,然后利用线性谱聚类对选择出的区域进行分割,最后通过分析避雷器热像的特征,实现避雷器故障的检测。实验结果说明所提出的算法可以有效地检测避雷器故障。