A SYSTEMATIC REVIEW ON RECENT ADVANCEMENTS IN DEEP AND MACHINE LEARNING BASED DETECTION AND CLASSIFICATION OF ACUTE LYMPHOBLASTIC LEUKEMIA

Abstract : Automatic Leukemia or blood cancer detection is a challenging job and is very much required in healthcare centers. It has a significant role in early diagnosis and treatment planning. Leukemia is a hematological disorder that starts from the bone marrow and affects white blood cells (WBCs). Microscopic analysis of WBCs is a preferred approach for an early detection of Leukemia since it is cost-effective and less painful. Very few literature reviews have been done to demonstrate a comprehensive analysis of deep and machine learning-based Acute Lymphoblastic Leukemia (ALL) detection. This article presents a systematic review of the recent advancements in this knowledge domain. Here, various artificial intelligence-based ALL detection approaches are in a systematic manner with merits and demits. The review of these schemes is conducted in a structured manner. For this purpose, segmentation schemes are broadly categorized into signal and image processing-based techniques, conventional machine learning-based techniques, and deep learning-based techniquesConventional machine learning-based ALL classification approaches are categorized into supervised and unsupervised machine learning is presented.
 EXISTING SYSTEM :
 ? The dropout layer is used to avoid overfitting. In 2020, Jha and Dutta presented a Chronological Sine Cosine Algorithm (SCA)-based deep CNN model to classify ALL successfully. Shahin et al. have proposed a new CNN-based framework (WBCsNet) to classify WBC efficiently. Claro et al. have suggested a deep-learning network (Alert-Net). It consists of 5 convolution layers, two fully connected layers, and a softmax layer. Conventional CNN needs a huge dataset to achieve outstanding performance. Since publicly available standard medical datasets are in small size, it is unable to properly fine-tune the weights to extract more important target-specific features efficiently. Thus, it yields relatively poor performance.
 DISADVANTAGE :
 ? Various researchers have employed the Watershed algorithm to segment overlapped and touched cells However, it faces an over-segmentation problem due to cells’ irregular shape. Some researchers have applied a marker-based watershed algorithm to segment overlapped and touched cells more effectively and solve the above problem . In this technique, we have first extracted the marker, and then the marker is imposed on the gradient to get the segmented cells They treat the segmentation as a pixel classification problem and segment the nucleus, cytoplasm, and background quite effectively
 PROPOSED SYSTEM :
 This article presents a brief analysis of recent advancements in deep and machine learning-based detection and classification of ALL. We have analyzed various existing methods of segmentation, feature extraction, and classification, which are employed to detect ALL efficiently. From this review, we also observed that among classical machine learning schemes, unsupervised schemes are preferred for segmentation tasks, whereas supervised schemes are preferred for classification tasks. More advanced deep learning or transfer learning-based systems can be designed by modifying the existing efficient models to make the system more accurate and faster as well
 ADVANTAGE :
 ? The ALL classification performance can be improved by employing an efficient ensemble transfer learning classifier that combines the benefits of all the transfer learning schemes used in this classifier. Hence, it enhances the diversity ability of feature learning, resulting in the extraction of more significant features. In addition, the combined benefits also improve the overall classification performances. In both these review works, the authors have emphasized the discussion of several segmentation approaches, whereas they have not focused on classification schemes
Download DOC Download PPT

We have more than 145000 Documents , PPT and Research Papers

Have a question ?

Mail us : info@nibode.com