Optimised Feature Selection for Identification of Carcinogenic Leukocytes Using Weighted Aggregation Based Transposition PSO

Abstract

In this paper, an automated leukocyte identification and classification methodology using microscopic blood smear images is proposed. After performing a colour based clustering for the identification of leukocytes, a large set of features are derived from the identified leukocytes. Thereafter a weighted aggregation-based transposition particle swarm optimisation technique (WATPSO) is proposed to select an optimised feature subset which is sufficient enough for classifying carcinogenic blast cells and normal cells efficiently. The proposed method is applied to a benchmark acute lymphoblastic leukaemia (ALL) data set and 97.30% test accuracy is achieved using only 6 optimal features. Experimental results demonstrate the usefulness of the proposed methodology in terms of high test accuracy and less number of informative selected features. The proposed method can be used as a decision support structure for haematologists for early diagnostic prediction of ALL.

KEYWORDS:

• Acute lymphoblastic leukaemia (ALL)
• Leukocytes
• Optimized feature subset
• Weighted aggregation-based transposition PSO (WATPSO)

Supplemental data

Supplemental data for this article can be accessed at https://doi.org/10.1080/03772063.2019.1682076.

Additional information

Notes on contributors

Subhajit Kar

Subhajit Kar received the BTech and MTech degrees in electrical engineering (with specialisation in power electronics and drives) from West Bengal University of Technology, Kolkata, India, in 2006, and from Bharath University, Chennai, India, in 2008. Currently, he is an assistant professor in the Department of Electrical Engineering at Future Institute of Engineering & Management, Kolkata, India. He is a member of IEEE (USA). His research interest includes computational biology, biomedical imaging, stochastic optimisation and machine learning techniques. Corresponding author. Email: [email protected]

Kaushik Das Sharma

Kaushik Das Sharma received the BTech and MTech degrees in electrical engineering from University of Calcutta, India, in 2001, 2004, respectively and PhD degree from Jadavpur University, India, in 2012. He is currently serving as an associate professor in University of Calcutta, India. He is a recipient of the Kanodia Research Scholarship in 2002 and University Gold Medal in 2004 from University of Calcutta. He was invited in University of Paris-Est Creteil, France as Teacher-Researcher in 2019. His research interests include computational intelligence, machine learning, robotics, etc. He has authored/coauthored about 54 technical articles, including 25 international journal papers. Email: [email protected]

Madhubanti Maitra

Madhubanti Maitra received the BE and ME degrees in electrical engineering (with specialisation in control systems) and the PhD degree in mobile computing from Jadavpur University, Calcutta, India, in 1989, 1991, and 2005, respectively. She is currently a professor with the Electrical Engineering Department, Jadavpur University. To her credit she has more than 70 international journal and conference papers. She acted as PI and Co-PI of several UGC and AICTE funded projects. Her research interests include control systems, mobile computing, and digital signal processing. Email: [email protected]