Multimodel Image Segmentation and Classification by MAP based graph cut and Improved VGG16
Jathin desan

Jathin desan*, 11th grade, in independence school, friscos texas, united states.

Manuscript received on April 11, 2020. | Revised Manuscript received on May 15, 2020. | Manuscript published on June 30, 2020. | PP: 465-472 | Volume-9 Issue-5, June 2020. | Retrieval Number: D7472049420/2020©BEIESP | DOI: 10.35940/ijeat.D7472.069520
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (

Abstract: Diagnosing brain diseases possess various inbuilt complexities to the nature of the diagnostic process. Brain tumor, Stroke, and Hemorrhage are the commonly prevailing disease and comprise more complexity in diagnosing where there arises the confusion in case of high grade or low-grade tumor and acute or sub-acute stroke. In general most of the prevailing algorithms is suited for the predicted of the image only employing the MRI or CT image. The paper mainly focused on the employment of a suitable proposed algorithm to adopt both the CT and MRI images for precise segmentation and classification. The segmentation algorithm is a map (map a posterior) based graph cut method The segmentation results are compared with the existing methods like (FCM) Fuzzy C Means and KFCM Kernel Fuzzy C Means and it is proved that our proposed system outperformed to the performance metrics. An improved VGG 16architecture is proposed for efficient classification. The overall classification results proved to be more efficient when compared with the existing R-CNN and NS-CNN methods. The paper focused on overcoming the difficulty and make a clear understanding of segmenting and classification irrespective of the nature of the diagnostic process.
Keywords: Brain tumor, Hemorrhage, Stroke, VGG 16 architecture, Recurrent Convolutional Neural Network, Graph curt based segmentation, MAP etc.