We study the problem of reconstructing binary images from four projections data in a fuzzy environment. Given the uncertainly projections, we want to find a binary image that respects as best as possible these projections. We provide an iterative algorithm based on fuzzy integer programming and linear membership functions.