Journal Publications

I sense overeating: Motif-based machine learning framework to detect overeating using wrist-worn sensing. S. Zhang, W. Stogin, and N. Alshurafa. Information Fusion, Volume 41, May 2018. [pdf]

Highlight: Authors face the important issue to detect people’s overeating activity, which often causes health issues, by using wrist-worn sensors. Specifically, a machine learning framework was designed to detect and accurately count the number of feeding gestures during an eating episode to characterize each eating episode with a feeding gesture count using a 6-axis inertial wrist-worn sensor. Results show optimal result for participant-dependent models. Generalization relies on a motif-based time-point fusion classification framework. Finally, mapping feeding gesture count to caloric intake, allowed to roughly estimate whether participants overate while watching television.

Conference Proceedings

Estimating Caloric Intake in Bedridden Hospital Patients with Audio and Neck-worn Sensors. S. Zhang, D. Nguyen, G. Zhang, R. Xu, N. Maglaveras, N. Alshurafa. The Third IEEE/ACM International Conference on Connected Health: Applications, Systems and Engineering Technologies (CHASE 2018).

Highlight: We present an approach for estimating calorie intake given a limited number of foods provided to patients in an in-bed setting. Data collected from a proximity sensor, inertial measurement unit, ambient light, and audio sensor placed around the neck are used to classify food-type consumed by second. We conducted a user study in a patient simulated lab setting, where 10 participants were asked to eat while sitting on a patient bed. Our system shows promise in distinguishing food items and predicting calorie intake in a bedridden participant setting given a limited set of food items.

When generalized eating detection machine learning models fail in the field. S. Zhang, R. Alharbi, M. Nicholson, and N. Alshurafa. Proceedings of the 2017 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2017 ACM International Symposium on Wearable Computers (Ubicomp 2017). [pdf]

Highlight: In this paper we use a wrist-worn sensor to test a generalized machine learning models’ reliability in detecting eating episodes through data processing. We process data from a 6-axis inertial sensor. Since most eating episodes do not occur while moving, we filter out periods of physical activity, and then use an advanced motif-based time-point fusion technique to detect feeding gestures. We also cluster each of the false alarms into four categories in an effort to identify the main behaviors that confound feeding gesture detection. This result shows clear challenges in characterizing eating episodes by using a single inertial-based wrist-worn sensor.

Food watch: detecting and characterizing eating episodes through feeding gestures. (Best Paper Award) S. Zhang, R. Alharbi, W. Stogin, M. Pourhomayun, B. Spring, and N. Alshurafa. Proceedings of the 11th EAI International Conference on Body Area Networks (BodyNets 2016). [pdf]

Highlight: To detect overeating passively, a machine learning framework was designed to detect and accurately count the number of feeding gestures during an eating episode to characterize each eating episode with a feeding gesture count. In this paper, our objective is to show the potential of commercial smartwatches to be used in detection of eating episodes with short durations confounded by other activities of daily living in order to truly capture all eating episodes in the field The effect of time-series segmentation and sensing configuration on the accuracy of detecting and characterizing feeding gestures is then analyzed. Finally, the effects of personalized and generalized machine learning models in predicting feeding gestures are compared.

An iterative dimensionality-scaling system for real-time health monitoring applications. H. Kalantarian, M. Sarrafzadeh, S. Zhang, and N. Alshurafa. 2016 IEEE International Conference on Healthcare Informatics (ICHI). [pdf]

Highlight: Wearable health-monitoring systems must achieve a balance between the often opposing goals of hardware overhead and classification accuracy. Prior works have presented various approaches to dynamically scale the accuracy of these systems as a function of available resources. In this paper, we present a framework which retroactively improves the accuracy of prior estimates when resources become available, using a novel global cost minimization function. We benchmark our algorithm on an audio-based nutrition monitoring dataset. Results confirm the efficacy of our technique.

Machine learning algorithms applied to detect feeding gestures S. Zhang, R. Alharbi, W. Stogin, K. Moran, AF. Pfammatter, B. Spring, and N. Alshurafa. The Obesity Society 2016. [pdf]

Auto-commissioning and adaptive tuning of servo control parameters in an electro-hydraulic system based on physical plant model. X. Cheng, S. Zhang, Y. Chen, H. Zhang, X. Zhu. 2014 IEEE 15th Workshop on Control and Modeling for Power Electronics (COMPEL). [pdf]

Highlight: In industrial servo control systems, the system performance is significantly affected by the physical parameter variance from different machine configuration and from the hydraulic loop non-linearity. Parameter detuning deteriorates the system performance and may even cause system instability issue, and parameter tuning requires much labor effort. In this paper, based on the physical plant model, an auto-commissioning and adaptive tuning method is proposed in a servo motor driven hydraulic control system. The hydraulic loop parameter which is time-variant is off-line or on-line estimated and then used for auto-commissioning and adaptive tuning of control parameters. The proposed method is verified with experiment to achieve high repeatability of dynamic and steady-state performance in different working conditions, and requires zero effort for parameter tuning.


If you would like to contact me, please email me on .