Motor and non-motor symptoms, such as hand tremors and palmar hyperhidrosis, provide critical diagnostic criteria for neurodegenerative diseases, such as Parkinson's disease. However, because conventional methods have limited ability to diagnose the early symptoms of these diseases, more sophisticated alternative approaches need to be developed. Here, we demonstrate a self-powered and touch-free image sensor (TFIS), fabricated with semiconductor InN nanowires (NWs), to monitor hand tremor and palmar hyperhidrosis in real time. The results suggest that the TFIS is potentially suitable for the early diagnosis of the symptoms of neurodegenerative diseases. An array composed of 16 TFIS chips is shown capable of detecting hand tremors, even very subtle movements that are clinically difficult to distinguish, in non-contact mode by leveraging the electrostatic induction in InN NWs caused by the external triboelectric effect. The array generates high output voltages in excess of 1.61 V without requiring an external power supply or signal processing, thus enabling the direct and intuitive recognition of physiological states. As an extension of its diagnostic utility, the system also detects palm moisture levels. The sensor system we developed provides a non-contact and quantitative sensing platform for early diagnosis and physiological signal-based interfaces in neurodegenerative disease management.