Light scattering poses significant challenges for biomedical optical imaging techniques. Diffuse scattering scrambles wavefront information, confounding easy analysis of signals reflected from or transmitted through biological tissues. For optical imaging techniques that employ only unscattered light components, the penetration depth is severely limited. In this thesis, we develop and discuss two general methods for dealing with large levels of light scattering in tissue. The first involves optimization of the signal-to-noise ratio (SNR) of coherence domain optical tomography techniques. The majority of the signal measured in these techniques is singly scattered. Thus, an improvement in SNR will improve the penetration depth of the system by picking out the weak signal contribution from increasing depths that would otherwise be buried in noise. We show that the SNR can be optimized in terms of image reconstruction algorithms, and in terms of detection parameters. An important detection parameter, the integration time, determines the dominant noise source of the measurement, and can be varied to obtain the maximal SNR. A second general method that will be discussed involves the time-reversal of scattered light components in tissues through the process of optical phase conjugation (OPC). OPC has long been used to remove optical aberrations and distortions, but has never before been applied to light scattering in tissues. We show that we are capable of time reversing light scattering in both chicken tissue sections and tissue phantoms, and characterize both the amplitude and resolution trends of the process. Finally, we provide the first successful results of OPC in living tissues.