Emotion Categorization from Video-Frame Images Using a Novel Sequential Voting Technique

Emotion categorization can be the process of identifying different emotions in humans based on their facial expressions. It requires time and sometimes it is hard for human classifiers to agree with each other about an emotion category of a facial expression. However, machine learning classifiers have done well in classifying different emotions and have widely been used in recent years to facilitate the task of emotion categorization. Much research on emotion video databases uses a few frames from when emotion is expressed at peak to classify emotion, which might not give a good classification accuracy when predicting frames where the emotion is less intense. In this paper, using the CK+ emotion dataset as an example, we use more frames to analyze emotion from mid and peak frame images and compared our results to a method using fewer peak frames. Furthermore, we propose an approach based on sequential voting and apply it to more frames of the CK+ database. Our approach resulted in up to 85.9% accuracy for the mid frames and overall accuracy of 96.5% for the CK+ database compared with the accuracy of 73.4% and 93.8% from existing techniques.

Emotion Categorization from Video-Frame Images Using a Novel Sequential Voting Technique

Emotion categorization can be the process of identifying different emotions in humans based on their facial expressions. It requires time and sometimes it is hard for human classifiers to agree with each other about an emotion category of a facial expression. However, machine learning classifiers have done well in classifying different emotions and have widely been used in recent years to facilitate the task of emotion categorization. Much research on emotion video databases uses a few frames from when emotion is expressed at peak to classify emotion, which might not give a good classification accuracy when predicting frames where the emotion is less intense. In this paper, using the CK+ emotion dataset as an example, we use more frames to analyze emotion from mid and peak frame images and compared our results to a method using fewer peak frames. Furthermore, we propose an approach based on sequential voting and apply it to more frames of the CK+ database. Our approach resulted in up to 85.9% accuracy for the mid frames and overall accuracy of 96.5% for the CK+ database compared with the accuracy of 73.4% and 93.8% from existing techniques.

Synchronization of walking “in the wild”

Marker-less video-based tracking promises to allow us to do movement science on existing video databases. We revisited the old question of how people synchronize their walking using real world data. We thus applied pose estimation to 348 video segments extracted from YouTube videos of people walking in cities. As in previous, more constrained, research, we find a tendency for pairs of people to walk in phase or in anti-phase with each other. Large video databases, along with pose-tracking algorithms, promise answers to many movement questions without experimentally acquiring new data.

More than Data: A Multivocal Inquiry into Video-Based Research on Learning and Teaching

Purpose This commentary aims to echo Wilkinson, Bailey, and Maher's (this volume) arguments about the affordances of videos and video databases in studying learning and teaching. Design/Approach/Methods This article illustrates a multivocal approach to the videos from the Video Mosaic Collaborative (VMC). In particular, three mathematics teachers in Shanghai were invited to watch and discuss a set of VMC videos. Two recurring themes concerning mathematics learning and teaching were identified in this video-cued interview and discussed in relation to the VMC Analytics. Findings The VMC videos played a mediating and facilitating role in the interview, helping the teachers notice and reflect on the mundane, implicit culture practices. Based upon this analysis, I argue that to tap into the potential of video in educational research, we need to see videos as more than data and look for more possibilities of using them. Originality/Value To open and further research dialogues, this article discusses future directions of using videos in educational research and serves as an invitation to creative explorations, in-depth conversations, ethical reflections, and cross-cultural collaborations on the use of videos in education.

Video Copyright Detection Using High Level Objects in Video Clip

Latest advancements in online video databases have caused a huge violation of copyright material misuse. Usually a video clip having a proper copyright is available in online video databases like YouTube without permission of the owner. It remains available until the owner takes a notice and requests to the website manager to remove copyright material. The problem with this approach is that usually the copyright material is downloaded and watched illegally during the period of upload and subsequent removal on request of the owner. This study aims at presenting an automatic content based system to detect any copyright violation in online video clips. In this technique, a video clip is needed from original video that is used to query from online video to find out shot similarity based on high level objects like shapes.