Recognizing emotions during social interactions has many potential applications with the popularization of low-cost mobile sensors, but a challenge remains with the lack of naturalistic affective interaction data. Most existing emotion datasets do not support studying idiosyncratic emotions arising in the wild as they were collected in constrained environments. Therefore, studying emotions in the context of social interactions requires a novel dataset, and K-EmoCon is such a multimodal dataset with comprehensive annotations of continuous emotions during naturalistic conversations. The dataset contains multimodal measurements, including audiovisual recordings, EEG, and peripheral physiological signals, acquired with off-the-shelf devices from 16 sessions of approximately 10-minute long paired debates on a social issue. Distinct from previous datasets, it includes emotion annotations from all three available perspectives: self, debate partner, and external observers. Raters annotated emotional displays at intervals of every 5 seconds while viewing the debate footage, in terms of arousal-valence and 18 additional categorical emotions. The resulting K-EmoCon is the first publicly available emotion dataset accommodating the multiperspective assessment of emotions during social interactions.

Introduction: With the COVID-19 outbreak, South Korea has been making contact trace data public to help people self-check if they have been in contact with a person infected with the coronavirus. Despite its benefits in suppressing the spread of the virus, publicizing contact trace data raises concerns about individuals’ privacy. In view of this tug-of-war between one’s privacy and public safety, this work aims to deepen the understanding of privacy risks of contact trace data disclosure practices in South Korea. Method: In this study, publicly available contact trace data of 970 confirmed patients were collected from seven metropolitan cities in South Korea (20th Jan-20th Apr 2020). Then, an ordinal scale of relative privacy risk levels was introduced for evaluation, and the assessment was performed on the personal information included in the contact trace data such as demographics, significant places, sensitive information, social relationships, and routine behaviors. In addition, variance of privacy risk levels was examined across regions and over time to check for differences in policy implementation. Results: It was found that most of the contact trace data showed the gender and age of the patients. In addition, it disclosed significant places (home/work) ranging across different levels of privacy risks in over 70% of the cases. Inference on sensitive information (hobby, religion) was made possible, and 48.7% of the cases exposed the patient’s social relationships. In terms of regional differences, a considerable discrepancy was found in the privacy risk for each category. Despite the recent release of government guidelines on data disclosure, its effects were still limited to a few factors (e.g., workplaces, routine behaviors). Discussion: Privacy risk assessment showed evidence of superfluous information disclosure in the current practice. This study discusses the role of "identifiability" in contact tracing to provide new directions for minimizing disclosure of privacy infringing information. Analysis of real-world data can offer potential stakeholders, such as researchers, service developers, and government officials with practical protocols/guidelines in publicizing information of patients and design implications for future systems (e.g., automatic privacy sensitivity checking) to strike a balance between one’s privacy and the public benefits with data disclosure.

The automated recognition of human emotions plays an important role in developing machines with emotional intelligence. Major research efforts are dedicated to the development of emotion recognition methods. However, most of the affective computing models are based on images, audio, videos and brain signals. Literature lacks works that focus on utilizing only peripheral signals for emotion recognition (ER), which can be ideally implemented in daily life settings. Therefore, in this paper we present a framework for ER on the arousal and valence space, based on using multi-modal peripheral signals. The data used in this work were collected during a debate between two people using wearable devices. The emotions of the participants were rated by multiple raters and converted into classes in correspondence to the arousal and valence space. The use of a dynamic threshold for ratings conversion was investigated. An ER model is proposed that uses a Long Short-Term Memory (LSTM)-based architecture for classification. The model uses heart rate (HR), temperature (T), and electrodermal activity (EDA) signals as its inputs with emotional cues. Additionally, a post-processing prediction mechanism is introduced to enhance the recognition performance. The model is implemented to study the use of individual and different combinations of the peripheral signals, as well as utilizing annotations from different ratings. Additionally, it is employed for classification of valence and arousal in an independent and combined fashion, under subject dependent and independent scenarios. The experimental results have justified the efficient performance of the proposed framework, achieving classification accuracy $>$ 96% and $>$ 93% for the independent and combined classification scenarios, accordingly. The comparison of the achieved performance against the baseline methods shows the superiority of the proposed framework and the ability to recognize arousal-valance levels with high accuracy from peripheral signals, in real-life scenarios.