How do I write a C++ program to read the JPG format picture, square the dots per pixel and output the image?

You probably don’t actually want to do what you are asking how to do. As Jim Clack pointed out in his answer, a JPEG file doesn’t store a raster image, it stores a compressed image as a set of transforms. When you convert this image into a bitmap in memory, that’s a process called rendering. Say I had a JPEG image, and the image was, in theory, “naturally” 400x400 pixels in size. I could render it into a bitmap in memory as a 400x400 image, and that would get me a 400x400 raster image, as a bitmap. And then I take that image, and do as you suggest, to magnify it into an 800x800 image, turning each pixel into a square composed of 4 pixels, at two pixels on a side. You’re going to get a pretty crappy image, and unless your intent is to make the image look as if it were being displayed on an old-time 8-bit computer, like an Atari 2600, or a Commodore 64, or an old game console, such as the original Nintendo/Famicon system, or an old Mattel Intellivision, etc. — you’re going to end up with a pretty terrible image. If, instead, you created an 800x800 canvas (an empty bitmap), and rendered the JPEG into it at that resolution directly — then the math is going to work out in such a way as to make the image look a whole heck of a lot better. The reason for this is that the 2x2 pixel regions won’t all be the same color; instead, what gets rendered into an individual pixel is going to depend on the adjacent information — in other words, it’s going to decompress into different pixel colors, for each of the pixels in the 2x2 region. read less