Deep learning approach for bone marrow cell detection and classification on whole-slide images

• Main Authors: Fadhil Abbas, Najwa, Shaizadi Meraj, Syeda, Zeki, Akram M., Shah, Asadullah
• Format: Proceeding Paper
• Language: English; English
• Published: IEEE 2023
• Subjects: T10.5 Communication of technical information
• Online Access: http://irep.iium.edu.my/109232/1/109232_Deep_Learning_Approach_for_Bone_Marrow_Cell_Detection_and_Classification_on_Whole-Slide.pdf; http://irep.iium.edu.my/109232/7/109232_Deep%20Learning%20Approach%20for%20Bone%20Marrow%20Cell%20Detection%20and%20Classification%20on%20Whole-Slide%20Images_SCOPUS.pdf; http://irep.iium.edu.my/109232/; https://ieeexplore.ieee.org/document/10346515

the bone marrow cell analysis is taken as the critical standard for diagnosing leukemia. However, owing to the diverse morphology of these bone marrow cells, a lot of patience along with extensive experience is required for the examination. In this research paper, a deep learning method has been proposed for intelligent detection and classification of the bone marrow cells through applying the object detection model and pattern recognition in order to minimize the error probability and work intensity as well as improve the work proficiency on contrary to the human recognition methods for bone marrow cell detection. The proposed method has used Faster R-CNN along with the generalized average precision loss (G-AP loss) method to improve the accuracy of the cell detection.