Multimodal Defensive Communication Database (DefComm-DB)

<p><strong>Description:&nbsp;</strong>DefComm-DB&nbsp;comprises 261 genuine <em>real-world </em>dialogues between English-speaking individuals in &#39;real-world&#39; settings that feature one of the defensive behaviours outlined in Birkenbihl&#39;s model of communication failures&nbsp;<a href="#biblio1">[1]</a>:</p> <ol> <li>Attacking the conversation partner (class <em>Attack</em>): videos that depict individuals actively attacking verbally, blaming the other person, or shifting the other person&#39;s attention to themselves.</li> <li>Withdrawing from the communication (class <em>Flight</em>): videos where people refuse to respond, withdraw from the conversation or change the topic or focus.</li> <li>Making oneself greater (class <em>Greater</em>): videos that depict individuals boasting, self-justifying in an aggressive manner, denying accusations, exhibiting a sense of dominance or superiority, or expressing indignation.</li> <li>Making oneself smaller (class <em>Smaller</em>): videos that display individuals engaging in self-deprecation, self-blame, exhibiting a sense of guilt, apologising, and expressing feelings of vulnerability or worthlessness.</li> </ol> <p>[1] Birkenbihl, V. (2013). Kommunikationstraining: Zwischenmenschliche Beziehungen erfolgreich gestalten. Schritte 1&ndash;6. : mvg Verlag.</p> <p>Key statistics on the dataset are provided in&nbsp;<a href="#tab:defcomm-stats">Table 1</a>. <strong>DefComm-DB</strong> features a variety of video topics, including interviews with celebrities and professional athletes, political debates, legal trials, TV shows, and video footage obtained by paparazzi, among others. The situations, number of participants, gender, age, and ethnicity vary from scene to scene.</p> <p>From each video, we retrieve audio, visual, and textual modalities. In this paper, we focus on the audio modality and the speech transcriptions.</p> <table> <caption><strong>Table 1: Statistics on DefComm-DB:</strong> number of video clips, mean duration (&mu;), standard deviation (&sigma;), minimum, maximum, and total duration of collected videos per class.</caption> <tbody> <tr> <th>Label</th> <th># video clips</th> <th>&mu; [s]</th> <th>&sigma; [s]</th> <th>min [s]</th> <th>max [s]</th> <th>&Sigma; duration [s]</th> </tr> <tr> <td>Attack</td> <td>112</td> <td>8</td> <td>9</td> <td>2</td> <td>46</td> <td>949</td> </tr> <tr> <td>Flight</td> <td>57</td> <td>9</td> <td>8</td> <td>2</td> <td>62</td> <td>494</td> </tr> <tr> <td>Greater</td> <td>45</td> <td>9</td> <td>6</td> <td>2</td> <td>25</td> <td>416</td> </tr> <tr> <td>Smaller</td> <td>47</td> <td>12</td> <td>8</td> <td>3</td> <td>49</td> <td>556</td> </tr> <tr> <td><strong>Total</strong></td> <td><strong>261</strong></td> <td><strong>9</strong></td> <td><strong>8</strong></td> <td><strong>2</strong></td> <td><strong>62</strong></td> <td><strong>2415</strong></td> </tr> </tbody> </table> <p>&nbsp;</p> <p>If you use <strong>DefComm-DB&nbsp;</strong>in your research&nbsp;work, you are kindly asked to acknowledge it in your publications.</p> <p><em>S. Amiriparian, L. Christ, R. Kushtanova, M. Gerczuk, A. Teynor, and B. Schuller, &ldquo;Speech-Based Classification of Defensive Communication: A Novel Dataset and Results,&rdquo; in Proc. INTERSPEECH 2023, Dublin, Ireland, ISCA, 8 2023. 5 pages, to appear.</em></p> <p><em>@inproceedings{Amiriparian23-SCO,&nbsp;</em></p> <p><em>author = {Shahin Amiriparian and Lukas Christ and Regina Kushtanova and Maurice Gerczuk and Alexandra Teynor and&nbsp; Bj&ouml;rn Schuller},&nbsp;<br> title = {{Speech-Based Classification of Defensive Communication: A Novel Dataset and Results}},&nbsp;<br> booktitle = {{Proceedings INTERSPEECH 2023}},&nbsp;<br> year = {2023},&nbsp;<br> address = {Dublin, Ireland},<br> publisher = {ISCA}, &nbsp;<br> month = {8},&nbsp;<br> note = {5 pages, to appear},<br> }</em></p>