A method includes receiving video data representative of a video frame. The video frame includes a first representation of a first object. The method also includes determining whether to replace the first representation in the video frame with a second representation of a second object based on an object matching rule. In response to determining that the first representation is to be replaced with the second representation, the method includes manipulating a three-dimensional model of the second object to generate the second representation of the second object so that a first visual property of the second representation matches a second visual property of the first representation. The method further includes generating a modified version of the video frame that includes the second representation in place of the first representation.