A common side effect of traumatic brain injury (TBI) is an expansion of the lateral ventricles. This side effect is mainly caused by two phenomena; The first is the accumulation of cerebrospinal fluid (CSF) in the ventricles, known as Post-Traumatic Hydrocephalus (PTH), and the second is cerebral atrophy, resulting in an enlarged appearance of the ventricles in a CT scan. Distinguishing between these two factors is clinically difficult since they could not be easily separable from CT scans. Nevertheless, this distinction must be made since it leads to different medical decisions. Patients suffering from PTH should be treated with shunt surgery. However, patients falsely diagnosed with PTH will not benefit from this surgery and will be exposed to a life-threatening procedure. Therefore, the diagnosis must be made with high confidence. Today, it is made by physicians based on the patient's medical record and CT scan. This makes the process vulnerable to human errors. Since computers are more precise and consistent than humans, over the past years more and more automated computational systems are integrated into medicine. We suggest that the decision whether to perform a shunt surgery on a patient suffering from ventricles expansion after TBI, can and should be done with the aid of a computerized system. In this work, we address this challange using image processing and machine learning techniques. Our research is conducted in collaboration with Dr. Sacher, rehabilitation expert and the head of the head-trauma department at Beit-Loewenstein rehabilitation institution. Analyzing the brain-scans of 30 patients, our results suggest that such solution is indeed feasible. Bigger data set may enable better learning, resulting in a more applicable and reliable system to aid physicians in making this decision.