Abstract:
Medical imaging data in today’s healthcare information systems is an essential part of diagnostics. Securing medical imaging data plays a critical role in the current time but today it is a complex task of maintaining data privacy so the ultimate intention of this study is to solve this problem.
In this project firstly we secure and compress the MRI images of the brain using cryptography and Huffman compression. AES (Progressed Encryption Standard) could be a symmetric-key encryption calculation that’s broadly utilized for scrambling information. It employments a piece cipher to scramble information in fine-tuned-size pieces of 128 bits and bolsters key sizes of 128, 192, and 256 bits. A lossless information compression calculation is huffman coding. The concept is to distribute variable length codes to input characters, with the code length decided by the recurrence of the coordinating character.
In this whole process, input images are encrypted and compressed and after that images are decrypted using AES algorithm and supplied as input to the pre-trained convolutional neural networks such as Alex-net and mobile net. The Alex net comprises 25 layers such as convolutional, batch-normalization, ReLU and max-pooling etc. The classification between the tumor and healthy images has been performed using the softmax layer. MobileNet is built with depth-separable convolutional layers. Each depthwise distinguishable convolutional layer comprises of a depthwise convolution and a pointwise convolution. Tallying the depthwise and pointwise convolutions as partitioned layers, MobileNet has 28 layers.
The execution of the proposed show has been endeavored on the publically accessible Kaggle dataset and a comparison is made between relegation calculations. The Alex net accomplished 97% of the precision whereas mobile net accomplished 99% of the precision that’s distant way better as compared to the most recent distributed investigative work in this space.
