Dears, I wish to measure the entropy of visual data such as image and video. I did transform these image and videos into frequency domain such as DCT, then measured the entropy with these pdfs for each AC components of image and videos. However, when I measured the entropy over spatial domain such as pixel, this entropy is differ to the entropy from DCT domain. What is the relationship between the entropy of spatial and frequency domains? (Such as Parseval's theorem) I couln't find this relation in the text book for signal processing or the google. Please give a comments for this relation. Thanks!