Attractive significance image base medicinal picture study and how to improve detection of brain tumours MRI be achievement thought inside current period outstanding towards augmented require of competent with accurate reports using semantic segmentation. In this paper researchers are tend use All convolution network method in replacement of FULL CONVOLUTION NETWORK and to detect brain tumor more effectively and accurately we are going to use a combination of DEEP LEARNING and MACHINE LEARNING which will make detection in a better way. The ALL CONVOLUTION NETWORK will use CONV which are designed for a modern convolution network. In the case of assessing and identifying the tumor instead of using 2d detection and dice cut here we are going to use 3d segmentation in detection which makes it more accurate. Different algorithms are work best for different sub regions and fusing several best algorithms will yield a good result in full segmentation with the help of FCN. And FCN to CONV layer replacement means great reduction in the number of parameters. It's cool to save the memory, but it's loss of flexibility nevertheless. In my experiments, CIFAR-10 classification accuracy dropped slightly after this single change, though it was certainly within one standard deviation. But I'm sure there are cases when this loss of flexibility makes a bigger impact.