Understanding and replicating human hand shape and grasping behavior are essential for improving physical human-robot interaction. In this work, we introduce a novel method for estimating both human hand geometry and grasping style using a tactile sensory humanoid hand. Our system integrates a silicone-based glove embedded with pressure sensors and mounted on a robotic hand, allowing users to perform naturalistic grasping gestures. By analyzing the tactile feedback generated during the interaction, we trained AI models to estimate individual hand shapes and classify grasping behaviors. A user study with 19 participants evaluated the comfort and usability of the system. Participants highlighted the softness and responsiveness of the glove. Feedback was used to identify key design improvements, including hand scaling, sensor distribution, and enhanced realism in tactile textures. Our results demonstrate the feasibility of using a tactile humanoid hand as an interactive tool for capturing nuanced human grasp style and hand size with 68% and 96%, respectively.