Geometrical Analysis of White Blood Cells by Using Image Segmentation Algorithm

Abstract

The technique behind image analysis is based on machine learning. In order for the software to recognize specific irregularities, it doesn\'t matter what field of medicine (or other industry) we\'re thinking about. The way AI learns differs dramatically from how humans learn. It makes use of enormous image databases that the software must interpret and analyze. In the present work, geometrical features of White Blood Cells such Area, Perimeter, Orientation, and Equivalent Diameter are calculated by using Image Segmentation approach.

Introduction

I. INTRODUCTION

Using sample data or prior knowledge, machine learning is the process of programming computers to maximize a performance criterion. We have a model that has been developed up to a certain point, and learning is the application of a computer program to maximize the model's parameters using training data or prior knowledge. The model may be descriptive to learn from the data or predictive to make future predictions. How to build computer programs that gradually get better with experience is a topic of research in the field of machine learning [1-5]. Tens of millions of compounds need to go through a series of tests in order to manufacture or find a new medicine, which is an expensive and time-consuming process. One and only one could produce a usable medication.

The lengthy multi-step process can be sped up in one or more of these processes using machine learning [6-10]. The field of personalized medicine has enormous development potential in the future, and machine learning may be key in identifying the types of genetic markers and genes that are responsive to a given therapy or medication. As it allows for improved disease assessment, customized medication or treatment based on a person's health information combined with analytics is a popular study topic. Another data flood may be beneficial for treatment efficacy with the rise in sensor-integrated technology and mobile applications with advanced monitoring system and health-measurement capabilities. Health optimization is made possible by personalized care, which also lowers overall healthcare expenses [11-15].

Machine Learning also finds application in manufacturing sectors for defects identification and material properties optimization [16-22]. Medical professionals all across the world are inundated with data of every description, which they must gather, handle, and analyze.

Image Processing approach is finding a wide range of applications [31-35].  As individuals, they are limited in their abilities and prone to weariness, which is bad for both their health and their capability to care for patients. In the healthcare industry, medical photographs make up about 90% of the data.

The amount of data that needs to be analyzed is growing along with the demand for medical imaging. It's crucial to keep in mind that machine learning-based solutions can and, in some instances, already have outperformed human doctors in terms of diagnosis accuracy. Of course, there will always be a need for some degree of expert oversight. When dealing with photos that are defective, partial, or otherwise of poor quality, the software can occasionally suffer. At that point, a human doctor's expertise and the capacity to view an image without turning it to ones and zeros may be required [23-30].

II. EXPERIMENTAL PROCEDURE

In the recent work, the Python programming has been developed for image segmentation purpose which was executed on SPYDER platform. The approach to image segmentation is based on a topological interpretation of the boundaries of the image. The closure process aids in filling in tiny cracks or tiny dark spots on the foreground objects. Morphological dilatation fills in small holes in the objects and increases object visibility. The Euclidean distance formula is used to calculate the distance transform.

A distance matrix is created once these distance values are computed for each and every pixel in a picture. It serves as a watershed transform input.

Figure 1 shows the input image of White Blood Cells which will be subjected to the image segmentation algorithm.

II. RESULTS AND DISCUSSION

Picture segmentation entails breaking an image down into a number of pixel-rich sections that are each represented by a mask or labeled image. We can analyze only the key portions of an image by segmenting it rather than processing the whole thing. In contrast to classifiers, which typically use a single encoder network, most image segmentation models in computer vision use an encoder-decoder network.

Figure 2 shows the segmented image of the white blood cell image and Table 1 shows the geometrical parameters used to define these cells.

The encoder transforms the input into a latent space representation, which the decoder transforms into segment maps, or, more precisely, maps indicating the positions of each object in the image. One of the simplest techniques for segmenting images is thresholding, which establishes a threshold for categorizing pixels into two groups. The threshold value determines which pixels are set to 1 and which pixels are set to 0. Pixels with values below the threshold value are set to 0. Thus, the image undergoes a process known as binarization in which it is transformed into a binary map. When the difference in pixel values between the two target classes is quite great, image thresholding is highly helpful since it is simple to select an average value as the threshold. When binarizing a picture, threshold is frequently employed to enable the employment of additional, binary-only algorithms like contour detection and identification.

Table 1: Obtained Results for Image Segmentation

Conclusion

Medical digital image processing can increase the quality of the image and lessen the impact of noise. Images that have been processed can clearly communicate the image\'s pathological and medical information as well as the ailment that is being focused on. The use of digital images is essential on a daily basis. The handling of images with a computer is referred to as medical imaging processing. This processing entails a wide range of methods and actions, including image acquisition, archiving, presentation, and communication. The image is a function that denotes a measure of the properties of an observed sight, such as illumination or color. The advantages of digital photos include quick and inexpensive processing, simple transmission and storage, immediate quality assessment, many copies while maintaining quality, quick and inexpensive reproduction, and flexible manipulation. Digital photos\' drawbacks include copyright infringement, the inability to resize while maintaining quality, the necessity for high-capacity memory, and the requirement for a quicker processor for modification.

References

[1] Mitchell, T., Buchanan, B., DeJong, G., Dietterich, T., Rosenbloom, P. and Waibel, A., 1990. Machine learning. Annual review of computer science, 4(1), pp.417-433. [2] Jordan, M.I. and Mitchell, T.M., 2015. Machine learning: Trends, perspectives, and prospects. Science, 349(6245), pp.255-260. [3] Dietterich, T.G., 1990. Machine learning. Annual review of computer science, 4(1), pp.255-306. [4] Shavlik, J.W., Dietterich, T. and Dietterich, T.G. eds., 1990. Readings in machine learning. Morgan Kaufmann. [5] Mohri, M., Rostamizadeh, A. and Talwalkar, A., 2018. Foundations of machine learning. MIT press. [6] Dimitri, G.M. and Lió, P., 2017. DrugClust: a machine learning approach for drugs side effects prediction. Computational biology and chemistry, 68, pp.204-210. [7] Menden, M.P., Iorio, F., Garnett, M., McDermott, U., Benes, C.H., Ballester, P.J. and Saez-Rodriguez, J., 2013. Machine learning prediction of cancer cell sensitivity to drugs based on genomic and chemical properties. PLoS one, 8(4), p.e61318. [8] Eitrich, T., Kless, A., Druska, C., Meyer, W. and Grotendorst, J., 2007. Classification of highly unbalanced CYP450 data of drugs using cost sensitive machine learning techniques. Journal of chemical information and modeling, 47(1), pp.92-103. [9] Winkler, D.A., 2021. Use of artificial intelligence and machine learning for discovery of drugs for neglected tropical diseases. Frontiers in Chemistry, 9, p.614073. [10] Gerdes, H., Casado, P., Dokal, A., Hijazi, M., Akhtar, N., Osuntola, R., Rajeeve, V., Fitzgibbon, J., Travers, J., Britton, D. and Khorsandi, S., 2021. Drug ranking using machine learning systematically predicts the efficacy of anti-cancer drugs. Nature communications, 12(1), pp.1-15. [11] Char, D.S., Shah, N.H. and Magnus, D., 2018. Implementing machine learning in health care—addressing ethical challenges. The New England journal of medicine, 378(11), p.981. [12] Panch, T., Szolovits, P. and Atun, R., 2018. Artificial intelligence, machine learning and health systems. Journal of global health, 8(2). [13] Futoma, J., Simons, M., Panch, T., Doshi-Velez, F. and Celi, L.A., 2020. The myth of generalisability in clinical research and machine learning in health care. The Lancet Digital Health, 2(9), pp.e489-e492. [14] Jain, V. and Chatterjee, J.M., 2020. Machine learning with health care perspective. Cham: Springer, pp.1-415. [15] McDermott, M.B., Wang, S., Marinsek, N., Ranganath, R., Foschini, L. and Ghassemi, M., 2021. Reproducibility in machine learning for health research: Still a ways to go. Science Translational Medicine, 13(586), p.eabb1655. [16] Mishra, A., 2020. Artificial intelligence algorithms for the analysis of mechanical property of friction stir welded joints by using python programming. Welding Technology Review, 92(6), pp.7-16. [17] Mishra, A. and Morisetty, R., 2022. Determination of the Ultimate Tensile Strength (UTS) of friction stir welded similar AA6061 joints by using supervised machine learning based algorithms. Manufacturing Letters, 32, pp.83-86. [18] Mishra, A., 2021. Supervised machine learning algorithms to optimize the Ultimate Tensile Strength of friction stir welded aluminum alloy. Indian J. Eng, pp.122-133. [19] Mishra, A., Sefene, E.M. and Tsegaw, A.A., 2021. Process parameter optimization of Friction Stir Welding on 6061AA using Supervised Machine Learning Regression-based Algorithms. arXiv preprint arXiv:2109.00570. [20] Thapliyal, S. and Mishra, A., 2021. Machine learning classification-based approach for mechanical properties of friction stir welding of copper. Manufacturing Letters, 29, pp.52-55. [21] Mishra, A., 2020. Machine learning classification models for detection of the fracture location in dissimilar friction stir welded joint. Applied Engineering Letters. [22] Mishra, A. and Vats, A., 2021. Supervised machine learning classification algorithms for detection of fracture location in dissimilar friction stir welded joints. Frattura ed Integrità Strutturale, 15(58), pp.242-253. [23] Razzak, M.I., Naz, S. and Zaib, A., 2018. Deep learning for medical image processing: Overview, challenges and the future. Classification in BioApps, pp.323-350. [24] Liu, L., Chen, W., Nie, M., Zhang, F., Wang, Y., He, A., Wang, X. and Yan, G., 2016. iMAGE cloud: medical image processing as a service for regional healthcare in a hybrid cloud environment. Environmental health and preventive medicine, 21(6), pp.563-571. [25] Agrawal, S. and Jain, S.K., 2020. Medical text and image processing: applications, issues and challenges. In Machine Learning with Health Care Perspective (pp. 237-262). Springer, Cham. [26] Bengtsson, E., 2005, June. Computerized cell image processing in healthcare. In Proceedings of 7th International Workshop on Enterprise networking and Computing in Healthcare Industry, 2005. HEALTHCOM 2005. (pp. 11-17). IEEE. [27] Krupinski, E.A., Williams, M.B., Andriole, K., Strauss, K.J., Applegate, K., Wyatt, M., Bjork, S. and Seibert, J.A., 2007. Digital radiography image quality: image processing and display. Journal of the American College of Radiology, 4(6), pp.389-400. [28] Erden, F., Velipasalar, S., Alkar, A.Z. and Cetin, A.E., 2016. Sensors in assisted living: A survey of signal and image processing methods. IEEE Signal Processing Magazine, 33(2), pp.36-44. [29] Tamir, A., Jahan, C.S., Saif, M.S., Zaman, S.U., Islam, M.M., Khan, A.I., Fattah, S.A. and Shahnaz, C., 2017, December. Detection of anemia from image of the anterior conjunctiva of the eye by image processing and thresholding. In 2017 IEEE Region 10 Humanitarian Technology Conference (R10-HTC) (pp. 697-701). IEEE. [30] Kong, Z., Li, T., Luo, J. and Xu, S., 2019. Automatic tissue image segmentation based on image processing and deep learning. Journal of healthcare engineering, 2019. [31] Mishra, A. and Pathak, T., 2020. Estimation of grain size distribution of friction stir welded joint by using machine learning approach. ADCAIJ: Advances in Distributed Computing and Artificial Intelligence Journal, 10(1), pp.99-110. [32] Mishra, A. and Patti, A., 2021. Deep Convolutional Neural Network Modeling and Laplace Transformation Algorithm for the Analysis of Surface Quality of Friction Stir Welded Joints. [33] Aparna, P. and Kishore, P.V.V., 2019. Biometric?based efficient medical image watermarking in E?healthcare application. IET Image Processing, 13(3), pp.421-428. [34] Mishra, S., Tripathy, H.K. and Acharya, B., 2021. A precise analysis of deep learning for medical image processing. In Bio-inspired neurocomputing (pp. 25-41). Springer, Singapore. [35] Marwan, M., Kartit, A. and Ouahmane, H., 2018. Security enhancement in healthcare cloud using machine learning. Procedia Computer Science, 127, pp.388-397.

Copyright

Copyright © 2022 Saurabh Mishra, Archana Kumari. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.