Abstract: This paper proposes a robust algorithm to recover shape from shading information. This algorithm can estimate effectively illuminant direction, surface albedo, the azimuth of the illuminant and recover shape using shading information along image contours. The elevation of the illuminant and surface albedo are estimated from image statistics, taking into consideration the effect of self-shading. With the estimated reflectance map parameters, we compute the surface shape using a new procedure that implements the smoothness constraint by making the gradients of reconstructed intensity close to the gradients of the input image. The new algorithm is data driven, stable, and updates surface slope and height maps simultaneously, and significantly reduces the residual errors in irradiance and integral terms. Finaly a hierarchical implementation of the SFS algorithm is presented.