Learning an artificial neural network to discover bit-quad-based formulas to compute basic object properties.

•A single-layer artificial neural network-based method learns bit-quad-based formulas to calculate basic object properties.•New bit-quad-based formulas for computing binary image area, perimeter, and contact perimeter are proposed.•An iterative elimination process reduces the number of terms of the formulas by removing irrelevant bit-quads.•The learned formulas reduce by about half the bit-quads needed to calculate the area, perimeter, and contact perimeter.