Machine learning techniques are used to explore the performance of boosted top quark tagging, treating jets as images. Tagging performances are studied in both hadronic and leptonic channels, employing a convolutional neutral network (CNN) based technique along with boosted decision trees (BDT). This computer vision approach is also applied to distinguish between left and right polarized top quarks, and an experimentally measurable asymmetry variable is constructed to estimate the polarization. Results indicates that the CNN based classifier is more sensitive to top quark polarization than the standard kinematic variables. It is observed that the overall tagging performance in the leptonic channel is better than the hadronic case, and the former also serves as a better probe for studying polarization.