Stroke is one of the lethal diseases that has significant negative impact on an individual's life. To diagnose stroke, MRI images play an important role. A large number of images are being produced day by day such as MRI (Medical Resonance Imaging), CT (Computed Tomography) X-Ray images and many more. Machine Learning algorithms are less efficient and time-consuming in localization of such medical images. Object detection using deep learning can reduce the efforts and time required in screening and evaluation of these images. In the proposed paper, several approaches such as RCNN (Region-based Convolutional Neural-Network), Fast R-CNN (Fast Region-based Convolutional Neural Network), Faster R-CNN (Faster Region-based Convolutional Neural Network with Region proposal Network), YOLO (You Only Look Once), SSD (Single-Shot Multibox Detector) and Efficient-Det are listed which can be used for stroke localization and classification. Comparison of RCNN, Fast R-CNN, Faster R-CNN, YOLO, SSD and Efficient-Det with accuracy are also present in this paper. A Chart of the Data Set available for object detection is also considered in this paper. By The maP (Mean-Average Precision) and the accuracy of every single method, it is identified that the speed and accuracy need to poise.