Most of the hundreds of exoplanets identified in the past 15 years have been detected indirectly---by monitoring the planets' effect on their host star's light. Now astronomers are able to actually image wide-separation exoplanets from large ground-based observatories by overcoming the extremely large brightness contrast between the stars and their faint exoplanets. This is allowing us to study these objects in detail, launching the field of comparative exoplanetary science. I will describe several of the techniques and results associated with this extremely challenging task of "high-contrast imaging." I will further describe some of our efforts at Caltech, including "Project 1640", a high contrast imaging program based at the Palomar 5m Hale Telescope. If time allows, I will also touch on two other projects: the first one involves deep Keck interferometric imaging of very young stars in star-forming regions which are providing snapshots of the very early epochs of planet formation and evolution.