Fix incorrect bitwise operation in pixel color calculation

[antoncuranz]

Hey,

I have found that in ImageBMP.cpp, there is a problem with the extraction of colors from the palette array:

For files with 4bit and 8bit color depth, the palette array seems to store two 3bit colors per entry. Depending on the last bit of the color index, a different color should be extracted from it using a shift and an AND operation.

However, the shift operation seems to be done before the AND. Because of this, the lower 3bit color is always returned. I fixed this by adding an extra pair of parentheses.