InfoMus Lab
The EyesWeb Research Project
The EyesWeb research project aims at exploring and developing models of interaction by extending music language toward gesture and visual languages, with a particular focus on the understanding of affect and expressive content in gesture. For example, in EyesWeb we aim at developing methods able to distinguish the different expressive content from two instances of the same movement pattern, e.g., two performances of the same dance fragment. Our research addresses the fields of KANSEI Information Processing and of analysis and synthesis of expressiveness in movement
 |
| A sample running EyesWeb application showing a visual representations of simple parameters extracted by movement. |
The figure above shows a sample running EyesWeb application, showing a visual representations of simple parameters extracted by movement using the EyesWeb Motion Analysis Library (available from version 2.5 of EyesWeb) and the upcoming EyesWeb Expressive Gesture Processing Library. Two examples of analysis are shown in the two windows: in the left window all the postures where peaks of velocity occurred are shown. The right window shows how space has been utilized in the last few seconds (blue area) and the current posture (white). The above application processes data in real time from a single videocamera or a video file.
 |
| An example of analysis of expressive cues, using the Silhouette Motion Images (SMI) technique. |
The above picture shows another simple example of analysis of expressive cues, using the Silhouette Motion Images (SMI) technique (see the papers in the publications section for further information). The window on the right shows a 2D map with the two axes corresponding to two motion features: fluentness (X axis) and quantity of movement (Y axis). High energetic and fluent movements cause a shift to the up-right part in the window (as it is in the snapshot of the figure above). Nervous/sudden energetic movements move the point to the up-left part. The current position in the map can be used to generate a suitable audio or visual output.