Advancing Early Detection: Dcnn for Automated Blood Cancer Diagnosis and Anomaly Detection

Authors: ,

Abstract: In today's modern healthcare system, accurate evaluation and diagnosis of blood cancer-related diseases continue to be of utmost importance, but they are also difficult to achieve due to the time-consuming manual analysis methods. Recent developments in computational methods, in particular those pertaining to machine learning and deep learning, have shown that they have the potential to significantly simplify this process. However, the lack of accurate and reliable automated tools for studying changes in blood cells is still a problem that slows down diagnostic procedures and makes early detection less accurate. The goal of this study is to show an advanced hybrid ensemble deep learning model that can automatically find and classify abnormal blood cells with a focus on finding leukaemia early. The model uses architectures like InceptionV3 and DenseNet201 and has stages for preprocessing, segmenting, augmenting, and classifying data. We achieve this by using a systematic framework. We meticulously classified 3,242 blood cell images into benign and malignant subtypes using the dataset. We also enhanced the dataset to increase its robustness. The model surpasses conventional methods by achieving an exceptional classification accuracy of over 99%. Using advanced visualisation tools, like Grad-CAM, also gives us a better understanding of how the model makes decisions. The methodology that has been proposed shows a tremendous deal of promise in terms of improving early detection and preventive diagnostics, which will ultimately contribute to timely medical interventions for diseases related to blood cancer.

Proof, ISSN / E-ISSN: 2944-9162 / 2732-9941, Volume 5, 2025, Art. #1

PDF
• HTML
DOI
XML
Cite
×
Citation Tools
WSEAS AMA
V. Raji, E. Rajendran, "Advancing Early Detection: Dcnn for Automated Blood Cancer Diagnosis and Anomaly Detection," Proof, vol. 5, pp. 1-9, 2025, DOI:10.37394/232020.2025.5.1
V. Raji, E. Rajendran. Advancing Early Detection: Dcnn for Automated Blood Cancer Diagnosis and Anomaly Detection. Proof. 2025;5:1-9. 10.37394/232020.2025.5.1
Copy to Clipboard
Citation copied to Clipboard

Certification

---