Public restroom detection on mobile phone via active probing

Abstract

Although there are clear benefits to automatic image capture services by wearable devices, image capture sometimes happens in sensitive spaces where camera use is not appropriate. In this paper, we tackle this problem by focusing on detecting when the user of a wearable device is located in a specific type of private space-the public restroom-so that the image capture can be disabled. We present an infrastructure-independent method that uses just the microphone and the speaker on a commodity mobile phone. Our method actively probes the environment by playing a 0.1 seconds sine wave sweep sound and then analyzes the impulse response (IR) by extracting MFCCs features. These features are then used to train an SVM model. Our evaluation results show that we can train a general restroom model which is able to recognize new restrooms. We demonstrate that this approach works on different phone hardware. Furthermore, the volume levels, occupancy and presence of other sounds do not affect recognition in significant ways. We discuss three types of errors that the prediction model has and evaluate two proposed smoothing algorithms for improving recognition.

Deep Room Recognition Using Inaudible Echos

“…The phone emits eight audible multi-tone chirps that cover a frequency range from 0.5 kHz to 4 kHz. In [16], the mel-frequency cepstral coefficients (MFCC) [41] of the acoustic echos triggered by audible sine sweep chirps are used to detect whether the phone's environment is a restroom. In [34], active acoustic sensing, combined with other passive sensing using magnetometer and barometer, classifies the phone's environment into six semantic locations: desk, restroom, meeting room, elevator, smoking area, and cafeteria.…”

“…The minimum needed volume of training data for our room recognition system will be investigated in Section 6. Existing studies on active acoustic sensing often use sine sweep chirp [16], Maximum Length Sequence [31], and multi-tone chirp [24] that cover a wide acoustic spectrum, including the audible range, to increase the information carried by the echos about the measured rooms. However, the audible chirps are annoying.…”

Pervasive Data Science

“…Bystander privacy is one that is often raised (and is thus the target of considerable research to date [1], [8]), but many more subtle data challenges also emerge. For example, the issue of provenance discussed above becomes important in ensuring that digital memory is an accurate representation of what occurred, particularly since the very need for augmented memory means the human may be unable to determine this for themselves.…”

Collecting Shared Experiences through Lifelogging: Lessons Learned

“…Discussions with long-term lifeloggers [2] indicate that the likelihood of violating the rules of private space does not reduce with experience and hence technical solutions (e.g. [7]) may be required to provide bystanders with the levels of privacy that they might reasonably expect. Reflecting on our previously described wearable camera image sample however, we note that only 0.60% of our captured images could actually be considered to violate privacy by depicting private areas with an additional 0.30% featuring bystanders/non-participants.…”

“…For situations where privacy risks can be clearly defined, automatic discontinuation of lifelogging capture seems like a promising solution (e.g. [7]), and for less clearly defined privacy criteria manual (user) in-situ control over capture may offer greater flexibility. Unsurprisingly Hoyle et al [11]'s recent study suggested that users preferred in-situ marking of photos for deletion rather than post-hoc processing and removal.…”