One of the important tasks in computer vision is the computation of object depth from acquired images. This paper explains the use of neural networks to calculate the depth of general objects using only two images, one of them being a focused image and the other one a blurred image. Having computed the power spectra of each image, they are divided to obtain a result which is independent from the image content. The result is then used for training Multi-Layer Perceptron (MLP) neural network (NN) trained by the backpropagation algorithm to determine the distance of the object from the camera lens. Experimental results are presented to validate the proposed approach. © Springer-Verlag Berlin Heidelberg 2006.