Today’s mobile devices have changed the way of interaction with computers. The touchscreen has to fulfill all requirements that prior were fulfilled by input devices like mouse and keyboard. These offer multiple levels of interaction through different mouse buttons and modifier keys. Instead, the input on touchscreens is flat. Additional dimensions like the touch duration extend the input vector but decrease input speed. Other options like movements gestures or input with multiple fingers are less accurate and require more space. To enable multiple levels of input on the touch screen we take the finger type as an additional dimension into account. We envision, that this is faster and more intuitive for users. Related work shows a wide range of possibilities in hand and gesture detection, but little methods on finger recognition. We enrich previous work by investigating a method for fully mobile prototypes. Touchscreens in today’s mobile devices are not able to detect the finger types, while the finger touches the screen. Hence, we build our own prototype by combining a tablet with a camera. We iteratively detected image features and trained finger recognition. We created a model based on data we collected in an experiment. Within 20 frames our method detects up to 72% accuracy of our participants’ hands and recognizes the correct finger with 70% accuracy.