Identification of Acute Lymphoblastic Leukemia in Microscopic Blood Image Using Image Processing and Machine Learning Algorithms
01 Sep 2018-pp 2359-2363
TL;DR: This proposed system uses openCV and skimage for image processing to extract relevant features from blood image and not just sheer number of features and further classification is carried out using various classifiers: CNN, FNN, SVM and KNN of which CNN gives the highest accuracy.
Abstract: Cancer has been plaguing the society for a long time and still there is no certain treatment; especially if detected in later stages. That is why early detection and treatment of cancer is of utmost importance. Acute lymphoblastic leukemia is a type of blood cancer which is known to progress very rapidly and prove fatal if there is a delay in detection. Detection of this type of cancer is carried out manually by observing the blood samples of patient under microscope and conducting various other tests. This process may produce undesirable drawbacks: slowness, nonstandardized accuracy since it depends on examiner's / pathologist's capabilities and fatigue due to work overload can cause human errors in detection. A few automated systems for detection of Acute Lymphoblastic Leukemia (ALL) have been proposed which involve extracting features from blood images using MATLAB and implementing different classifiers to produce results, which gave remarkable accuracies though not enough for practical usage. Our proposed system is further improving the classification accuracy. It uses openCV and skimage for image processing to extract relevant features from blood image and not just sheer number of features and further classification is carried out using various classifiers: CNN, FNN, SVM and KNN of which CNN gives the highest accuracy of 98.33%. CNN and FNN are written using TensorFlow framework. The accuracies obtained by other classifiers: FNN, SVM, and KNN are 95.40%, 91.40% and 93.30% respectively.
Citations
More filters
[...]
TL;DR: Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis.
Abstract: Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).
13,246 citations
TL;DR: This study shows the systematic review of promising research areas and applications of DL models in medical diagnosis and medical healthcare systems, and the prevalent DL models, their architectures, and related pros, cons are discussed to clarify their prospects.
Abstract: With time, AI technologies have matured well and resonated in various domains of applied sciences and engineering. The sub-domains of AI, machine learning (ML), deep learning (DL), and associated statistical tools are getting more attention. Therefore, various machine learning models are being created to take advantage of the data available and accomplish tasks, such as automatic prediction, classification, clustering, segmentation and anomaly detection, etc. Tasks like classification need labeled data used to train the models to achieve a reliable accuracy. This study shows the systematic review of promising research areas and applications of DL models in medical diagnosis and medical healthcare systems. The prevalent DL models, their architectures, and related pros, cons are discussed to clarify their prospects. Many deep learning networks have been useful in the field of medical image processing for prognosis and diagnosis of life-threatening ailments (e.g., breast cancer, lung cancer, and brain tumor, etc.), which stand as an error-prone and tedious task for doctors and specialists when performed manually. Medical images are processed using these DL methods to solve various tasks like prediction, segmentation, and classification with accuracy bypassing human abilities. However, the current DL models have some limitations that encourage the researchers to seek further improvement.
65 citations
TL;DR: It is found that the best pretrained models are VGG and GoogleNet compared to AlexNet by achieving 100% accuracy for training and testing and VGG also proven to have better result based on the training graph which is more stable and contains less error compared to the other two models.
Abstract: Acute Lymphoblastic Leukemia (ALL) is a disease that is defined by uncontrollable growth of malignant and immature White Blood Cells (WBCs) which is called lymphoblast. Traditionally, lymphoblast analysis is done manually and highly dependent on the pathologist’s skill and experience which sometimes yields inaccurate result. For that reason, in this project an algorithm to automatically detect WBC and subsequently examine ALL disease using Convolutional Neural Network (CNN) is proposed. Several pretrained CNN models which are VGG, GoogleNet and Alexnet were analaysed to compare its performance for differentiating lymphoblast and non-lymphoblast cells from IDB database. The tuning is done by experimenting the convolution layer, pooling layer and fully connected layer. Technically, 70% of the images are used for training and another 30% for testing. From the experiments, it is found that the best pretrained models are VGG and GoogleNet compared to AlexNet by achieving 100% accuracy for training. As for testing, VGG obtained the highest performance which is 99.13% accuracy. Apart from that, VGG also proven to have better result based on the training graph which is more stable and contains less error compared to the other two models.
13 citations
Cites background or result from "Identification of Acute Lymphoblast..."
...It is also widely known by its ability to cater image related problems [22]....
[...]
...Other than that, there are also research on ALL identification which compares the result of machine learning and CNN and CNN showed the best performance result [22]....
[...]
16 Nov 2020
TL;DR: In this paper, the authors proposed a hands-on approach in detecting the irregular blood components (e.g., Neutrophils, Eosinophils, Basophils and Monocytes) that are typically found in a cancer patient.
Abstract: It is reported that since 2016 there are over sixty thousand diagnosed cases of Leukemia in the United States of America alone. It also suggests that Leukemia is the most common type of cancer seen in the age of twenty. Although the study is based on a Western country, it is equally alarming for an Asian country like Bangladesh where healthcare system is not up to the standard. Researches show that the Chronic Lymphocytic Leukemia has about 83% five-year long survival rates. This paper focuses on Acute Lymphocytic Leukemia (ALL) as this is the most common type of Leukemia in Bangladesh. It is common knowledge among oncologists, that cancer is much easier to treat if it is detected in the early stages. Thus the treatment needs to begin as early as possible. We propose a hands-on approach in detecting the irregular blood components (e.g., Neutrophils, Eosinophils, Basophils, Lymphocytes and Monocytes) that are typically found in a cancer patient. In this work, we first identify 14 attributes to prepare the dataset and determine 4 major attributes that play a significant role in determining a Leukemia patient. We have also collected 256 primary data from Leukemia patient. The data is then processed using microscope to obtain images and fetch into Faster-RCNN machine learning algorithm to predict the odds of cancer cells forming. Here we have applied two loss functions to both the RPN (Region Convolutional Neural Network) model and the classifier model to detect the similar blood object. After identifying the object, we have calculated the corresponding object and based on the count of the corresponding object we finally detect Leukemia. The mean average precision observed are 0.10, 0.16 and 0, where the epochs are 40, 60 and 120, respectively.
8 citations
TL;DR: This comprehensive review provides an in-depth analysis of the current AI practices in the field of hematology, with a special focus on the type ofHematologic malignancies and the patient’s cancer stage to determine future research directions in blood cancer.
Abstract: Background Machine learning (ML) and deep learning (DL) methods have recently garnered a great deal of attention in the field of cancer research by making a noticeable contribution to the growth of predictive medicine and modern oncological practices. Considerable focus has been particularly directed toward hematologic malignancies because of the complexity in detecting early symptoms. Many patients with blood cancer do not get properly diagnosed until their cancer has reached an advanced stage with limited treatment prospects. Hence, the state-of-the-art revolves around the latest artificial intelligence (AI) applications in hematology management. Objective This comprehensive review provides an in-depth analysis of the current AI practices in the field of hematology. Our objective is to explore the ML and DL applications in blood cancer research, with a special focus on the type of hematologic malignancies and the patient’s cancer stage to determine future research directions in blood cancer. Methods We searched a set of recognized databases (Scopus, Springer, and Web of Science) using a selected number of keywords. We included studies written in English and published between 2015 and 2021. For each study, we identified the ML and DL techniques used and highlighted the performance of each model. Results Using the aforementioned inclusion criteria, the search resulted in 567 papers, of which 144 were selected for review. Conclusions The current literature suggests that the application of AI in the field of hematology has generated impressive results in the screening, diagnosis, and treatment stages. Nevertheless, optimizing the patient’s pathway to treatment requires a prior prediction of the malignancy based on the patient’s symptoms or blood records, which is an area that has still not been properly investigated.
7 citations
References
More filters
[...]
TL;DR: Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis.
Abstract: Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).
13,246 citations
"Identification of Acute Lymphoblast..." refers background in this paper
...Machine learning is a branch of artificial intelligence (AI) research that employs a variety of statistical, probabilistic and optimization tools to learn from past examples and to then use that prior training to classify new data, identify new patterns or predict novel trends [25]....
[...]
29 Dec 2011
TL;DR: A new public dataset of blood samples is proposed, specifically designed for the evaluation and the comparison of algorithms for segmentation and classification, to offer a new test tool to the image processing and pattern matching communities.
Abstract: The visual analysis of peripheral blood samples is an important test in the procedures for the diagnosis of leukemia. Automated systems based on artificial vision methods can speed up this operation and increase the accuracy and homogeneity of the response also in telemedicine applications. Unfortunately, there are not available public image datasets to test and compare such algorithms. In this paper, we propose a new public dataset of blood samples, specifically designed for the evaluation and the comparison of algorithms for segmentation and classification. For each image in the dataset, the classification of the cells is given, as well as a specific set of figures of merits to fairly compare the performances of different algorithms. This initiative aims to offer a new test tool to the image processing and pattern matching communities, direct to stimulating new studies in this important field of research.
369 citations
"Identification of Acute Lymphoblast..." refers background in this paper
...The description of such a public dataset and metrics for evaluation of performance of different classification algorithms are discussed [6]....
[...]
14 Jul 2004
TL;DR: This paper presents a methodology to achieve an automated detection and classification of leucocytes by microscope color images and firstly individuates in the blood image the leucocyte from the others blood cells, then it extracts morphological indexes and finally it classifies the leukocytes by a neural classifier in Basophil, Eosinophils, Lymphocyte, Monocyte and Neutrophil.
Abstract: The classification and the count of white blood cells in microscopy images allows the in vivo assessment of a wide range of important hematic pathologies (i.e., from presence of infections to leukemia). Nowadays, the morphological cell classification is typically made by experienced operators. Such a procedure presents undesirable drawbacks: slowness and it presents a not standardized accuracy since it depends on the operator's capabilities and tiredness. Only few attempts of partial/full automated systems based on image-processing systems are present in literature and they are still at prototype stage. This paper presents a methodology to achieve an automated detection and classification of leucocytes by microscope color images. The proposed system firstly individuates in the blood image the leucocytes from the others blood cells, then it extracts morphological indexes and finally it classifies the leucocytes by a neural classifier in Basophil, Eosinophil, Lymphocyte, Monocyte and Neutrophil.
193 citations
20 Jul 2005
TL;DR: The presented paper shows the effectiveness of an automatic morphological method to identify the Acute Lymphocytic Leukemia by peripheral blood microscope images.
Abstract: The early identification of acute lymphoblastic leukemia symptoms in patients can greatly increase the probability of recovery. Nowadays the leukemia disease can be identified by automatic specific tests such as Cytogenetics and Immunophenotyping and morphological cell classification made by experienced operators observing blood/marrow microscope images. Those methods are not included into large screening programs and are applied only when typical symptoms appears in normal blood analysis. The Cytogenetics and Immunophenotyping diagnostic methods are currently preferred for their great accuracy with respect to the method of blood cell observation which presents undesirable drawbacks: slowness and it presents a not standardized accuracy since it depends on the operator's capabilities and tiredness. Conversely, the morphological analysis just requires an image -not a blood sample- and hence is suitable for low-cost and remote diagnostic systems. The presented paper shows the effectiveness of an automatic morphological method to identify the Acute Lymphocytic Leukemia by peripheral blood microscope images. The proposed system firstly individuates in the blood image the leucocytes from the others blood cells, then it selects the lymphocyte cells (the ones interested by acute leukemia), it evaluates morphological indexes from those cells and finally it classifies the presence of the leukemia.
189 citations
Additional excerpts
...This automated system offers remarkable accuracy [2]....
[...]
TL;DR: The automated Leukaemia detection system analyses the microscopic image and overcomes the drawbacks of a manual method and is successfully implemented in MATLAB.
118 citations
"Identification of Acute Lymphoblast..." refers background in this paper
...5% but further classification of type of blasts is not shown [12]....
[...]