Microwave imaging is a new medical imaging modality that is used to estimate and track changes in the dielectric properties (i.e. permittivity and conductivity) of tissue. It has shown promise in clinical applications like breast imaging as it is comfortable, safe and low-cost. However, due to the complexity of breast tissue a signal can travel more than one path through different tissue to get from a transmitter to a receiving antenna. This phenomenon is called multipath propagation. This work explores the potential of using multipath information to improve the permittivity estimates in microwave breast imaging. To extract the multipath information from the signals, the Hilbert-Huang Transform (HHT) was used. Data from simple simulations and phantoms with inclusions were analyzed to understand the signal shapes and patterns around an inclusion using the HHT. From these scans, multipath signals were detectable from their distinct characteristics in the time-frequency domain. From there, the signals from more complex scenarios were analysed to ensure the HHT was still effective in a realistic environment. With all these findings a signal processing algorithm was developed and applied to simulation, phantom and human data. When applied to the simulation and phantom data, the algorithm proved to be better at determining the size and permittivity of inclusions when compared to 2D projections of the models. Finally, when applied to patient data, the algorithm demonstrated strong potential to create images with improved clinical information and was useful in tracking changes in the breast tissue over time. Overall, this work demonstrated the necessity of incorporating multipath information into the signal processing algorithms of microwave imaging systems.