It still doesn't matter. If you move the rectangle over the pixmap, it will force the pixmap to redraw as different areas are exposed. And if in the process of redrawing the pixmap you are also rescaling it from the camera image, that is a huge amount of work to do every few ms and it gets worse as the number of pixels (size of the pixmap) increases. Try what I suggested: use an overlay widget to draw the ROI so your ROI is not in the same scene (or even in the same graphics view) as the pixmap.Rectangle is not child of QGraphicsPixmapItem .This Rectangle and Image from camera(Which is i am drawing on QGraphicsPixmapItem) are childs of QGraphicsScene.
I would probably not even bother with the G-V architecture for the pixmap part at all - if the pixmap occupies the entire window, then just use a QWidget-based view to display it, and use the G-V architecture in the overlay widget only. If the scene coordinates match the pixmap dimensions, then all the coordinates will be self-consistent.
Bookmarks