A Comparative Approach of Brain Tumor Detection Using SVM, DCT and Huffman Coding in Compressed Domain

Background: This study does the qualitative analysis of segmentation technique with compression and authentication to provide improved and advanced e–health care system. This study addresses three main problems of brain tumor detection, one is to find the region of interest, second is to authenticate patient information and third is to minimize memory space for a large dataset of medical images. Methods: The proposed approach combines a region based segmentation using support vector machine, authentication using discrete cosine transformation and compression using Huffman coding. Storing image data is the major and longtime problem associated with images, especially when the sizes of the images are large. The size of the image is directly reflected with the storing size, more space it will require to be stored when the size of the image is large. The main aim of this work is to open a new way of diagnosing tumor detection under the compressed domain and so making suitability for mobile computing and internet based medical analysis. Discussion: The proposed algorithm combines in such a way that it will improve the performance parameters and also reduce the space for storage, without sacrificing the quality of pixels or image. Conclusion: To evaluate the performance of our proposed mechanism, we conducted simulations on different medical images. The simulation results prove the significance in terms of quality parameters and imperceptibility.