Sony Ericsson patent reads your face, builds a playlist

Sony Ericsson have patented a system whereby a cellphone could be controlled by facial expressions, shuffling between music tracks according to visual mood analysis.  The patent, titled "Generating music playlist based on facial expression", describes capturing an image of a user and using facial recognition to build a custom playlist depending on the listener's mood.

The system could either recognize moods by spotting particular facial characteristics or by comparing the whole face against presets; alternatively, a hybrid of the two might be applied. 

In a more complex implementation, Sony Ericsson suggest that brief video could be captured, tracking emotional states of happiness, sadness, surprise, anger, fear, disgust, neutrality, etc.  These would be compared to the musical characteristics of the tracks stored on the cellphone or PMP, similar to how Sony Ericsson's existing SenseMe application works.

"FIG. 8 [shown above] is a diagram illustrating the concepts described herein. As illustrated, David is listening to a Bruce Springsteen song on a mobile phone that includes a DMP. David is in a very happy mood because he recently found out that he won the lottery. As David adjusts the volume of the DMP, camera 260 may capture an image of David's face. Based on the captured image, FER component 420 may determine that the facial expression of David matches a category (e.g., a happy mood category) contained in category field 514 of FER database 500. Category field 514 may include a previously linked song associated with the happy mood category (e.g., a song by Metallica)." Sony Ericsson patent

A slightly more recent patent by the company, "Portable hands-free device with sensor", suggests other areas in which Sony Ericsson are exploring intelligent feature management based on how the user handles their device.  It presents functions – such as switching between stereo and mono sound – as passively controlled by capacitance, inductance, pressure, temperature, light, movement, or acoustical sensors. 

[via USEB]