We tested our approach using two co

We tested our approach using two complex transparent objects WATER GLASS and WINE GLASS. Figure 10shows the reconstruction result on a transparent WATER GLASS which has a complex shape. The result is acceptable in regions where salient highlights can be detected. Note the faithfulness of the recovered shape using our approach where true highlights are picked and false ones are suppressed. Figure 11 shows another result on a transparent WINE GLASS where a lot of orientation discontinuities are present. Note in the normal map the fine details of the wine glass preserved. Two views of the reconstructed surface, alongside with the real object in similar views are shown. Note that a similar glass was also used in fluorescent immersion range scanning [6]. Our surface generated using our shoestring budget looks reasonable and comparable in many ways.