Telescopes with unobstructed pupil are known to deliver clean point spread function (PSF) to their focal plane, in contrast to traditional telescopes with obstructed pupil. Recent progress in the manufacturing aspheric surfaces and mounting accuracy favors unobstructed telescopes over obstructed telescopes for science cases that demand stable and clean PSF over the entire field-of-view. In this paper we compare the image quality of an unobstructed Three-Mirror-Anastigmat (TMA) design with that of an obstructed TMA. Both the designs have the same primary mirror, effective focal length, field-of-view and detector characteristics. We demonstrate using simulated images of faint elliptical galaxies imaged through the two designs, that both the designs can measure morphological parameters with same precision, if the PSF is reconstructed within 12 arc-minutes of the source. We also demonstrate that, the unobstructed design delivers desirable precision even if the PSF is reconstructed 50 arc-minutes away from the source. Therefore the PSF of unobstructed design is uniform over a wider field-of-view compared to an obstructed design. The image quality is given by the 1$\sigma$ error-bars (68% confidence level) in the fitted values of the axis-ratio and position-angle of the simulated galaxies.