Rehabilitation plays a crucial role in the recovery process from injuries, offering pain relief and speeding up recovery. However, traditional rehabilitation methods often require in-person appointments with healthcare professionals, making them inefficient and costly. Encouraging patients to perform exercises at home is a potential solution, but without professional feedback, patients lack motivation and adherence. Even telerehabilitation, which employs information and communication technology, has limitations, including dependency on clinician availability. In recent years, the integration of Computer Vision (CV) has gained momentum in enhancing telerehabilitation. While earlier methods involved patient-robot interactions and marker-based CV systems, contemporary approaches favor markerless, equipment-free CV solutions driven by artificial intelligence. This paradigm shift holds promise for accessible and efficient rehabilitation, having the potential to revolutionize patient care. Our work aims to address a notable gap in the literature by consolidating recent research on accessible intelligent rehabilitation systems using smartphone cameras and CV and by compiling and analyzing the necessary literature for future studies.