| Sumario: | Background: Brain diagnosis process is an important medical activity; it can be diagnosis through many techniques. The numerous diagnosis processes are available for Brain scanning, such as CT, MRI, X-ray and CTA. If Medical applications are fast and real, then diagnosis process is easy; these can help the patient's life. Objective: In this work CT scan-based brain diagnosis system is proposed through CNN-GB technique. A real and accurate brain diagnosis process gives the better treatment with efficient success rate. Method: In this work the brain abnormalities have been identified through machine and deep learning mechanisms. This work requires dataset (Kaggle dataset) and real time CT scan images. CNN and GBML techniques are applied on CT brain images for identify the disorders. Results: This proposed method achieves 0.992 accuracy and 0.993 Tp these are good results compared to earlier method. Conclusion: In this investigation a CT scan-based brain diagnosis application is designed. This architecture is a combination of adaptive median filter, CNN deep learning model and gradient boosting machine learning. Due to combination of three algorithms getting the accurate results compared to earlier models. This application is achieving PSNR = 56.20, SSIM is 0.99, accuracy is 0.992, error rate is 0.05 average time 0.07 have achieved. Moreover, this method is verified on various high density of noise, at all type of densities implemented application attains the more improvement.
|
|---|