Pattern recognition and classification of two cancer cell lines by diffraction imaging at multiple pixel distances
Graphic abstract
Introduction
Cell classification by recognition of image patterns is of fundamental interest and can have wide applications in life science and clinics [1], [2], [3], [4], [5], [6], [7], [8]. Compared to microscopy, flow cytometry (FCM) with imaging capability allows rapid data acquisition and extraction of pattern parameters from large numbers of cells on a single-cell basis [9]. The high rate of image acquisition through FCM demands development of automated image analysis algorithms to process and analyze the big data of acquired images. Based on previous studies of coherent light scattering [10], [11], [12], [13], [14], [15], we have developed a polarization diffraction imaging flow cytometry (p-DIFC) method to measure the spatial distribution of light scattered by single cells illuminated by a linearly polarized laser beam [16], [17], [18], [19], [20], [21], [22]. Different from fluorescence imaging, the cross-polarized diffraction image (p-DI) pair data acquired by the p-DIFC method record intensity distribution of coherent light scattered by a cell due to the heterogeneous 3D distribution of intracellular refractive index and needs no cell staining. The acquired p-DI pair data are results of coherent superposition of wavefields emitted by the induced dipoles of molecules in the imaged cell and thus present speckle patterns carrying molecular and morphological information of the cell [12], [13], [15]. We have shown that the p-DI pair data can be used to accurately distinguish two cell lines of high morphological similarity derived from human T and B cancer cells and two types of prostate cells derived from cancer and normal cells [20], [22].
Conventional cell images are typically acquired with non-coherent fluorescent light through a microscope or imaging FCM which present 2D projections of a 3D object. Image segmentation is generally needed for further analysis [9]. In contrast, the speckle patterns in a p-DI pair result from superposition of coherent wavefields from all excited intracellular molecules as “digital holograms” and needs no segmentation. A pixel-based global image processing algorithm often suffices to quantify the patterns or textures of such images that can be automated for rapid processing. We have developed a gray-level-co-occurrence-matrix (GLCM) based software to quantitatively characterize textures of p-DI pair data [22], [23], [24], [25]. The GLCM parameters were determined from a matrix of elements given by the co-occurring probabilities of paired pixels separated by d as the displacement vector with |d|=1. A total of 38 parameters were obtained from each p-DI pair to form feature vectors, which were used to train support vector machine (SVM) classifiers with different kernels. With the above algorithms, we have shown that the p-DIFC method performs well for accurate and label-free cell classification through automated image pattern recognition.
Despite the attractive qualities of the p-DIFC method, however, we have found that the performances of the trained SVM classifiers are not consistent when they were applied to test sets of p-DI pairs data acquired in different runs of measurements. The causes of inconsistency may relate to the variation of diffraction patterns in different data sets, which are due to cell speed fluctuations leading to different degree of blurring for images acquired with conventional CCD cameras, positioning errors of cells relative to the focus of incident beam and the high sensitivity of extracted GLCM parameters on fine pattern changes among pairs of nearest-neighbor pixels that are unrelated to cell morphology or molecular composition. In our previous study of T versus B cell lines and PC3 versus PCS prostate cells [20], [22], SVM classifiers had to be trained and tested with the p-DI pairs acquired in the same measurement run for achieving high values of classification accuracy A. The value of A decreases markedly if the SVM classifier trained with data from one run was applied to the test data acquired in another run on a different day [22]. For example, the value of A reduces from 99.5% to 62.8% in the case of PC3 versus PCS cells. Such reduction in performance prevents application of the p-DIFC method with pre-trained SVM classifiers on p-DI data acquired later.
The current study focuses on pattern analysis of p-DI data and effect of blurring as a part of our research efforts to solve the accuracy-dropping problem. A new configuration of illumination and imaging has been developed to eliminate or significantly reduce motion blur with one time-delay-integration (TDI) CCD camera and decrease the sensitivity of image patterns of acquired data with enlarged focal spot for the incident beam [21]. With this imaging configuration we have investigated the GLCM approach of pattern analysis with d≥1 to improve the robustness of the SVM classifiers. In this report, we present the results of pattern recognition and classification on two cancer cell lines, HL-60 versus MCF-7, by the p-DI data acquired with the new imaging configuration. The dependence of p-DI parameters and classification accuracy on d has been analyzed to examine the benefits of using different values of d for improved performance. We have also blurred the measured p-DI data with no or little blurring by window smoothing to investigate the effect of blurring on classification. These results demonstrate that the new imaging configuration and GLCM analysis with different d can improve significantly the robustness of SVM based classification.
Section snippets
Related work
Automated image analysis for cell classification often proceed by extracting feature parameters related to cell morphology and/or image texture. Morphological parameters can be regarded as quantitative extension of human perception and obtained by characterizing geometric structures of organelles in images acquired by bright-field or fluorescent microscopy, which are typically performed with cells stained with contrast or fluorescent reagents. In deriving these parameters, image segmentation is
The p-DIFC system and cell measurement
Details of the p-DIFC system have been published elsewhere for cell positioning by hydrodynamic focusing in a square flow channel and imaging of coherent light scatter [16], [17], [18], [20], [21]. Briefly, a continuous-wave solid state laser (MGL-III-532-100, CNI) was used to produce an incident beam of 532 nm in wavelength. Two cylindrical lenses of 500 mm and 60 mm in focal lengths place the focus of the incident beam on the core fluid carrying the cells in an elliptical cross-section of major
Analysis of diffraction images
Three groups of measurements have been carried out in different days to acquire p-DI pairs from the HL-60 and MCF-7 cell suspension samples using a prototype system (DIFC2-P3, WavMed Technologies Corp.) equipped with one TDI camera as shown in Fig. 1(A). During each measurement, cell suspensions in volumes of about 0.1 mL were loaded multiple times into the core fluid tube at the top of the flow chamber. A loaded sample was driven by a syringe pump into the core fluid nozzle inside the flow
Discussion
By acquiring paired p-DI data from imaged cells, the p-DIFC method provides a means to perform image based pattern recognition and cell classification that is significantly different from conventional approaches of microscopy and FCM. As the result of coherent superposition of wavefields emitted by induced molecular dipoles, the p-DI pair data carry rich morphological and molecular information in the forms of complex diffraction patterns. To extract and to utilize the information, however, one
Conclusion
A cell classification study on HL-60 and MCF-7 cells has been carried out to assess the performance of pattern recognition and machine learning algorithms on p-DI data acquired in different days with a new imaging configuration using one TDI camera. The results demonstrate that the significant reduction of blurring in p-DI data and choice of different pixel distances can significantly improve the consistency of best performing SVM classifiers. Additional studies should be pursued to cell types
Conflict of interest
None declared.
Acknowledgments
We wish to thank Dr. Fu Zheng of Tianjin Medical University for preparing the HL-60 and MCF-7 cell samples and Dr. Wenhuan Jiang for helps on the development of classification software. X.H. Hu acknowledges grant support by the Tianjin Science and Technology Commission and Y. Feng acknowledges grant supports from National Natural Science Foundation of China (#81041107 and #81171342).
He Wang received his BS degree from the Shandong University in 2010 and PhD degree from the Tianjin University in 2016, both are in biomedical engineering. His current research interests include biological cell imaging and development of TDI cameras.
References (32)
- et al.
Biological shape characterization for automatic image recognition and diagnosis of protozoan parasites of the genus Eimeria
Pattern Recogn.
(2007) - et al.
An automated pattern recognition system for classifying indirect immunofluorescence images of HEp-2 cells and specimens
Pattern Recogn.
(2016) - et al.
Preserving boundaries for image texture segmentation using grey level co-occurring probabilities
Pattern Recogn.
(2006) - et al.
Nuclear patterns of human breast cancer cells during apoptosis: characterisation by fractal dimension and co-occurrence matrix statistics
Cell Tissue Res.
(2005) - et al.
Color graphs for automated cancer diagnosis and grading
IEEE Trans. Bio-Med. Eng.
(2010) - et al.
Feature extraction based on co-occurrence of adjacent local binary patterns
- S. Ghosh, V. Chaudhary, Feature analysis for automatic classification of HEp-2 florescence patterns: computer-aided...
- et al.
Benchmarking HEp-2 cells classification methods
IEEE Trans. Med. Imaging
(2013) - B.C. Lovell, G. Percannella, M. Vento, A. Wiliem, Performance evaluation of indirect immunofluorescence image analysis...
- et al.
Measurement of Cytoplasmic to Nuclear Translocation, Current Protocols in Cytometry
(2009)
Two-dimensional angular optical scattering for the characterization of airborne microparticles
Opt. Lett.
Angular distribution of light scattered by single biological cells and oriented particle agglomerates
Appl. Opt.
Simulations of light scattering from a biconcave red blood cell using the FDTD method
J. Biomed. Opt.
Effect of detailed cell structure on light scattering distribution: FDTD study of a B-cell with 3D structure constructed from confocal images
J. Quant. Spectrosc. Radiat. Transf.
A miniaturized wide-angle 2D cytometer
Cytom. A
Angle-resolved Mueller matrix study of light scattering by B-cells at three wavelengths of 442, 633 and 850 nm
J. Biomed. Opt.
Cited by (37)
Preferred vector machine for forest fire detection
2023, Pattern RecognitionSingle-detector dual-modality imaging flow cytometry for label-free cell analysis with machine learning
2023, Optics and Lasers in EngineeringStochastic configuration networks for self-blast state recognition of glass insulators with adaptive depth and multi-scale representation
2022, Information SciencesCitation Excerpt :Therefore, the recognition performance of the learning model deteriorates when faced with singular samples or similar samples belonged to different categories. From the perspective of pattern recognition [8,19,25], the man-made annotation pattern includes repeated comparison of visual features with multi-level and multi-scale, the reliability evaluation of self-blast recognition results in real-time, and adaptive regulation of the learning model according to the feature applicability of samples. In order to construct the generalization model of the self-blast state of glass insulators, it is of great significance to optimize the model from the perspective of pattern recognition.
Density peak clustering using global and local consistency adjustable manifold distance
2021, Information SciencesCitation Excerpt :Clustering is an unsupervised classification method which aims at dividing data into several homogeneous groups or clusters according to their similarities, so that data points within the same group are similar to each other while data points across different groups are dissimilar. Since clustering can effectively discover the inherent structure information in data and pave the way for further data analyses, it has been extensively used in various real-world applications such as pattern recognition [36,13], data mining [35,2], computer vision [31] and machine learning [28]. Up to now, a number of clustering algorithms have been developed, which can be mainly categorized into partition-based methods [16,25], hierarchical-based methods [9,10], grid-based methods[43,32], model-based methods [42,30], graph-based methods [34] and density-based methods [11] etc.
Polarization image fusion with self-learned fusion strategy
2021, Pattern RecognitionCitation Excerpt :Compared with conventional imaging technologies, polarization imaging provides not only intensity images but also polarization ones. Thus, polarization imaging has been widely used in military and civil applications [4–7], for example, mine detection, anti-camouflage recognition, dehazing, medical diagnosis, and scratch detection of transparent materials. In other fields such as infrared and visible images fusion, and multi-focus image fusion, lots of fusion methods have been proposed in the literature [8–11].
Self-blast state detection of glass insulators based on stochastic configuration networks and a feedback transfer learning mechanism
2020, Information SciencesCitation Excerpt :The detection of insulator self-blast states has been studied in various countries [30,33,34], but current detection models of insulator self-blast states based on image processing technology cannot meet the requirements of accuracy and speed. From the perspective of pattern recognition [10,24,26], the human inspection mode can be summarized as follows: based on rich experience, operators repeatedly compare visual features by adopting a multihierarchy and discrimination approach, evaluate the credibility of cognition results of multistate objects in real time, and discriminate the insulator self-blast states. Heterogeneous features with different hierarchies represent different analytical aspects and perspectives of cognitive objects.
He Wang received his BS degree from the Shandong University in 2010 and PhD degree from the Tianjin University in 2016, both are in biomedical engineering. His current research interests include biological cell imaging and development of TDI cameras.
Yuanming Feng is a professor of biomedical engineering at the Tianjin University. His research interests include diffraction imaging flow cytometry and measurement technique of radiation-induced apoptosis.
Yu Sa received his BS and Ph.D degrees from Tianjin University and he is a lecturer at Department of Biomedical Engineering of Tianjin University. His research interests include instrument development, diffraction imaging, and computational fluid dynamics modeling studies.
Jun Qing Lu received her BS/MS degrees in physics from Nankai University and Ph.D degree from University of California, Irvine. She is currently an associate professor at East Carolina University and mainly interested in numerical studies of light scattering in turbid media and with biological cells.
Junhua Ding is an associate professor of computer science with East Carolina University. He received his BS, MS and Ph.D, all in computer science, in 1994, 1997, and 2004, respectively. His research interests are in software engineering and data engineering, and published over 60 peer reviewed papers in these fields.
Jun Zhang received his BS degree in 2010 and PhD degree in 2016 from the Tianjin University, both are in biomedical engineering. His current research interests are in image analysis and medical physics.
Xin-Hua Hu received his BS and MS degrees from Nankai University, Tianjin, China, in 1982 and 1985, an MS degree in physics from Indiana University in 1986, and PhD degree in physics in 1991 from the University of California at Irvine. He joined the physics faculty in 1995 and is currently a professor at East Carolina University. His main research interests relate to the investigations of light scattering and their applications in probing tissues and cells.