Abstract

A brain Haemorrhage and Hematoma detection system has been designed and developed. This work presents a new approach to automated detection of haemorrhage and hematoma using k-means clustering technique that separates the lesion region from healthy tissues in computed tomography images. The computed tomography feature images used for the intracranial haemorrhage and the hematoma detection consists of non-enhanced computed tomography images. The application of the proposed technique for early detection of intracranial haemorrhage and the hematoma is demonstrated to improve efficiency and accuracy of clinical observations. The results are quantitatively evaluated by a human expert. The average overlap metric, average precision and average recall between the results obtained using the proposed approach and the ground truth are 0.85, 0.80 and 0.97 for haemorrhage and hematoma.

Paper Details