Medical applications have greatly benefited from the rapid progress in the field of computer vision. While previous studies primarily focused on analyzing diagnostic procedure data to predict the presence of diseases, there has been less attention given to areas such as patient behavior monitoring, and motor and mental disorder assessment. From a clinical perspective, behavior monitoring tools offer several key benefits: (i) they provide complementary, objective, and quantitative information to clinicians; (ii) they enable the detection and quantification of events that are challenging to observe, such as nocturnal falls; (iii) they reduce the time and effort required for documenting relevant diagnostic information; and (iv) they allow assessment in locations and clinics where human expertise may be limited. Vision-based systems have gained significant attention due to their non-invasive nature, demonstrating promising results in analyzing patient-specific poses and behaviors (including facial and body motions) across various clinical contexts, such as Epilepsy, Parkinson's disease, autism spectrum disorders, breathing disorders, infant motions, and pain management. Although these developments will never replace the expertise of individual clinicians, they can enhance medical decisions and ultimately improve the standard of care provided to patients by offering more quantitative evidence and appropriate decision support.