Abstract: A method for determining a location of a caption in a video signal associated with a Region Of Interest (ROI), such as a face or text, or an area of high motion activity. The video signal is processed to generate ROI location information, the ROI location information conveying the position of the ROI in at least one video frame. The position where a caption can be located within one or more frames of the video signal is then determined on the basis of the ROI location information. This is done by identifying at least two possible positions for the caption in the frame such that the placement of the caption in either one of the two positions will not mask the ROI. A selection is then made among the at least two possible positions. The position picked is the one that would typically be the closest to the ROI such as to create a visual association between the caption and the ROI.