{"id":390,"date":"2016-03-21T11:20:02","date_gmt":"2016-03-21T10:20:02","guid":{"rendered":"http:\/\/www.lustcon.de\/WordPress\/wearables-at-work\/?p=390"},"modified":"2016-05-02T08:57:08","modified_gmt":"2016-05-02T06:57:08","slug":"adl-challenge","status":"publish","type":"post","link":"https:\/\/www.lustcon.de\/WordPress\/wearables-at-work\/2016\/03\/adl-challenge\/","title":{"rendered":"ADL Detection Challenge at IEEE Healthcom 2016 Conference"},"content":{"rendered":"
At the 18th IEEE International Conference on E-health Networking, Application \u00a0& Services (IEEE Healthcom’16)<\/a><\/strong>\u00a0in Munich, September 14-17, 2016, an ADL detection challenge will be performed. The ADL (Activities of Daily Life<\/em>) detection challenge will be organized as a half-day workshop within the conference taking place at the BMW Welt.<\/a>\u00a0We kindly ask for you participation in the challenge described below.<\/p>\n <\/p>\n Automatic recognition and surveillance of the regular occurrence of ADLs \u2013 activities of daily life like tooth brushing, shaving, drinking, eating, combing, writing and more \u2013 can help in detecting accidents in the household or gradual changes in the behavior of elderly persons. In detail, the identification of health critical situations like dehydration, tumbles or disorientation is of great importance. If the elderly persons are suffering from mild cognitive impairment (MCI) of beginning dementia and a potentially dangerous situation is detected, external help can be alerted automatically in order to clarify the situation, to prevent further damages or mitigate them.<\/p>\n Smartwatches and wearables are ideal tools collecting necessary data with their embedded sensors and applying various recognition algorithms for detecting ADLs based on these data. The advantage of these wearables technologies is their \u201cinvisibility\u201d, meaning others will not immediately recognize that a wearer may have a handicap. This is in contrast to other products which typically stigmatize their users (e.g. emergency alert wrist bands resp. buttons).<\/p>\n This challenge is targeting to bring together developers, researchers and companies who want to demonstrate their achievements in this area.<\/p>\n A set of training data from real ADLs together with their classification will be made available through this web site. Interested participants can download the training data<\/a> after registration and create detection models for the data. Participants are completely free how to implement their detection task. The only restriction is that later on participants must be able to run their models in real time in the workshop.<\/p>\n In the workshop the organizers will distribute several test data, also from real ADLs. The participants can then use their trained models to run and classify the test data and prepare their classification results in a standardized way. They will have about 60 minutes for this task. After this the participants show and explain their models, solutions and classifications results in a short presentation. After the presentations the best solution is chosen and receives a prize. This is followed by a general discussion and roundup of the workshop.<\/p>\nMotivation<\/h4>\n
Organization of the Challenge<\/h4>\n
Time Table<\/h4>\n