In a previous study, we presented a new technique for representation of the shape of a scattering surface. A sensor based on two parallel fiber arrays yielded a source-detector intensity matrix (SDIM). In that study, it was shown that convex and concave polyacetal plastic (Delrin) surfaces could be accurately distinguished using the proposed technique. A simplified simulation model for calculating the SDIM was used, assuming that backscattered light was generated by Lambertian sources in the illuminated surface. These simulations showed discrepancies compared to measurements, probably due to the absence of light scattering in the model. Here, we will present an improved model, based on the Monte Carlo technique for light transport in turbid media. The optical properties of the Delrin phantoms were estimated by means of different measurement techniques. The optical properties and the geometry of the Delrin phantoms were implemented in the model along with the spatial distribution of the source and detector fibers of the sensor. The SDIM was extracted from backscattered photons exiting the turbid medium from the curved surface. The SDIM:s obtained with the Monte Carlo model, showed a much closer agreement with the measurements than those obtained with the Lambertian model. The small discrepancies observed are probably due to spatially varying optical properties of the plastic phantoms. Measurements, using the previously described sensor, of the SDIM from Delrin pieces with convex and concave surfaces, are compared to the SDIM extracted from simulations using the Monte Carlo model.