Prostate Brachytherapy has emerged as a common and effective treatment modality for early stage low risk prostate cancer, wherein around 75 small radioactive seeds are implanted into the prostate to eradicate the cancer by emitting radiation. The subject of this dissertation is the capability to intra-operatively monitor the 3D seed locations using 2D X-ray images and make treatment adjustments, promising to make a significant improvement in cancer control. We modeled the overall medical problem as a cascade of computer science problems, predominantly in image processing, pose estimation, minimum-weight multipartite graph matching and 3D image reconstruction. The primary contribution of this dissertation is a theoretical and computational solution using techniques like linear algebra, error analysis, combinatorial and non-linear optimization. The solution is consolidated into a simple software module and integrated with a commercial Brachytherapy installation. The contributions of the dissertation have been extensively validated using simulation and phantom experiments. Furthermore; the software system has been successfully employed intraoperatively, for treatment control in a phase-I trial with six patients. From a clinical perspective, the proposed solution does not alter the current surgical workflow, uses already existing infrastructure and is economically feasible.