Most of the text was positioned on top of the image/video elements, my translator was able to detect them.