Provided are systems and methods for forming personalized videos including a self-image of a user. An example method includes displaying a plurality of stock videos to a user, where a video of the plurality of stock videos is generated based on a live video featuring an actor wearing a mask and facing a video camera and the mask is a marker for insertion of a self-image of the user, providing an interface for selecting a stock video from the plurality of stock videos, and, upon determining that the user has selected the stock video, forming a personalized video using the selected stock video and the self-image of the user.