You already got a good link describing the differences.
When images are displayed by an application they have to be converted to bitmaps using the colour depth of the graphic system. Because this is usually 24-bit nowadays (8 bit per RGB channel), preparing the images with that colour depth is a good start. Converting a PNG image to a bitmap is really fast when the colour depth matches and the image has not to be scaled. A compressed image must always be decrompressed which tooks some time. But this should be no problem with common systems.
As already mentioned, scaling took some processing time too. While the times are again usually not a problem, proper scaling avoids artefacts and may reduce the image size.
- If necessary convert images to 24-bit colour depth (32-bit with alpha channel).
- Scale the images to the final resolution.
- Choose the format according to the tips from the link: JPEG for (very) large images like photos and PNG for small and medium size images (graphics).