Home > Articles > All Issues > 2025 > Volume 13, No. 3, 2025 >
JOIG 2025 Vol.13(3):275-285
doi: 10.18178/joig.13.3.275-285

Automatic Detection and Classification of Cerebral Microbleeds Using 3D CNN

M. Mohsin Jadoon 1,2,*, Victor Torres-Lopez 3, Sharjeel A. Butt 2,4, Santosh B. Murthy 5, Guido J. Falcone 3, and Seyedmehdi Payabvash 6
1. Pak-Austria Fachhochschule: Institute of Applied Sciences and Technology (PAF-IAST), School of Computing Sciences, Pakistan
2. Department of Electrical Engineering, International Islamic University Islamabad, Pakistan
3. Department of Neurology, Yale School of Medicine, New Haven, CT, USA
4. School of Automation, Central South University, Hunan, China
5. Department of Neurology, Weill Cornell Medicine, New York, NY, USA
6. Department of Radiology, Columbia University Medical Center, New York, NY, USA
Email: mohsin.khan@spcai.paf-iast.edu.pk (M.M.J.); victor.torreslopez@yale.edu (V.T.-L.); sharjeel401@csu.edu.cn (S.A.); sam9200@med.cornell.edu (S.B.M.); guido.falcone@yale.edu (G.J.F.); sp4479@columbia.edu (S.P.)
*Corresponding author

Manuscript received July 23, 2024; revised August 21, 2024; accepted October 16, 2024; published June 12, 2025.

Abstract— Cerebral Microbleeds (CMBs) are referred to tiny foci of hemorrhage in brain parenchyma which are smaller than 5 (to 10) mm in size. The presence of CMBs is implicated in pathophysiology of cognitive impairment, dementia, radiation-induced vascular injury, traumatic brain injury, hypertensive microangiopathy, and aging. On brain Magnetic Resonance Imaging (MRI) scans, CMBs appear as hypointense foci, most notable on T2*-weighted or Susceptibility-Weighted Imaging (SWI). Detecting these tiny microbleeds with naked eye is a difficult and time-consuming task for radiologists. In this study we developed an algorithm for automatic detection of CMBs. We applied a two-step strategy: at first, we applied pre-processed 2D image dataset to You Only Look Once (YOLO V2) for detection of CMBs. Then, these detected CMBs locations are used to segment 3D patches from their original SWI volume in the datasets. Next, these patches are used as inputs for Convolution Neural Network (CNN). In the second step, we reduced the number of False Positives (FP) and improved our classification accuracy using 3D CNN. We used two datasets consisting of 979 patients: 879 of whom for training of models, and the remainder for independent validation. We were able to achieve an accuracy of 81% and reduce the FPawg to 0.16.

Keywords—Cerebral Microbleeds (CMBs), classification, detection, You Only Look Once (YOLO), 3D Convolution Neural Network (CNN)

Cite: M. Mohsin Jadoon, Victor Torres-Lopez, Sharjeel A. Butt, Santosh B. Murthy, Guido J. Falcone, and Seyedmehdi Payabvash, "Automatic Detection and Classification of Cerebral Microbleeds Using 3D CNN," Journal of Image and Graphics, Vol. 13, No. 3, pp. 275-285, 2025.

Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC-BY-4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.

PREVIOUS PAPER
A Novel Deep Learning Approach for Speckle Denoising Using Hyperparameter Tuning
NEXT PAPER
Enhancing Pneumonia Classification Performance through CNN Architecture Optimization and Hyperparameter Tuning