Retrieving graphics object from image object

The reason for this is the color depth of the image that you're using to retrieve the Graphics object. The way color depth works is this. An image may have a palette, or not. If it has a palette, all pixels are indexes into the palette table and the palette table contains the actual colors. If no palette, the pixels represent their actual color. Images with no palette have more colors than those with palettes, but require more space. 24-bit color images would require up to 50 million bytes just to hold the palette, and 3 bytes per pixel to point to the location in the palette. Instead, it just uses 3 bytes per pixel to represent the color, and no palette at all. Images with 8-bit color only need 768 bytes for the palette, and one byte per pixel. Anyway, you just need to use a color depth without a palette (which means non-indexed) to obtain your Graphics object.