Continuous prediction of perceived traits and social dimension in space and time (2014)
Venue: | In Proc. of IEEE Int. Conf. on Image Processing |
Citations: | 1 - 1 self |
Citations
1606 | Video Google: A text retrieval approach to object matching in videos
- Sivic, Zisserman
- 2003
(Show Context)
Citation Context .... For each block, 4bin gradient and 5-bin optical flow histograms were computed and concatenated into a 162-length feature vector. Each clip was then represented by using Bag-of-Words (BoW) formalism =-=[17]-=-. BoW formalism provides a compact and rich description in terms of the number of local feature occurrences. It performs K-means clustering to find the representative cluster centers, and then convert... |
1414 | The elements of statistical learning: Data mining, inference, and prediction, 2nd ed
- Hastie, Tibshirani, et al.
- 2009
(Show Context)
Citation Context ...tained by calculating the average of time-interval-dependent histograms. Finally, we modelled the relationship between the BoW histograms and the aggregated scores by using Lasso and Ridge Regression =-=[18]-=-. A separate regression model per traitsocial dimension was trained by applying leave-one-subjectout cross-validation strategy. Each time the parameters of the regression model were optimized over the... |
735 | Learning realistic human actions from movies
- Laptev, Marszalek, et al.
- 2008
(Show Context)
Citation Context ...o obtaining better interpretations of the individual judgements. We address the prediction problem of perceived traitsocial dimensions by employing histogram of gradient and histogram of optical flow =-=[14]-=- in conjunction with a linear regression method. Continuous annotations are used in two separate frameworks: (i) continuous prediction in space (CPS) and (ii) continuous prediction in space and in tim... |
106 |
Toward a structure- and process-integrated view of personality: Traits as density distributions of states.
- Fleeson
- 2001
(Show Context)
Citation Context ...s might have concentrated more on the verbal content rather than the visual cues. This would require extending the work presented here to combine visual and audio features. Recent works in psychology =-=[21]-=- introduced the concept of personality state as a behavioral episode wherein a person behaves more/less extravertedly depending on the situation. A recent study on automatic prediction of perceived tr... |
88 |
la Torre. Supervised descent method and its applications to face alignment
- Xiong, De
- 2013
(Show Context)
Citation Context ...In other words, we first extracted separate features per frame and then mapped these features onto the rating values at their corresponding time instant. We used the facial landmark point detector of =-=[19]-=- which results in 49 landmark points per frame. From each landmark, we calculated HoG+HoF values and obtained 49× 162 = 7938 length feature vectors. Prior to the regression analysis, we reduced the di... |
39 |
The semaine database: Annotated multimodal records of emotionally colored conversations between a person and a limited agent.
- McKeown, Valstar, et al.
- 2012
(Show Context)
Citation Context ...ppear to be more static and stable over time while other dimensions appear to be more dynamic. 2. DATA AND ANNOTATION Data. In this work, we used the audio-visual recordings from the SEMAINE database =-=[15]-=-. This database contains recordings of human subjects interacting with virtual agents in a naturalistic scenario. We took into account 10 different subjects, each communicating with 3 semi-automatic S... |
8 |
Please, tell me about yourself: automatic personality assessment using short self-presentations
- Batrinca, Mana, et al.
- 2011
(Show Context)
Citation Context ...ions, namely, agreeableness, conscientiousness, extroversion, neuroticism and openness to experience. Although most of the existing literature focused on a subset of personality of the Big Five model =-=[1, 2, 3, 4]-=-, few studies also took into account other social dimensions such as likeability [5], group involvement and individual engagement [6], persuasiveness as well as its potential correlation with the five... |
8 |
FaceTube: Predicting Personality from Facial Expressions of Emotion in Online Conversational Video.
- Biel, Teijeiro-Mosquera, et al.
- 2012
(Show Context)
Citation Context ...gent had to perform a task on the computer screen and the agent could display different levels of collaboration, from agreeable and stable to less likely to compromise and neurotic. A number of works =-=[8, 3]-=- focused on “video blogs” and generated annotations through the crowdsourcing scheme. Biel et al. [8] detected facial expressions of emotions (anger, happiness, sadness, etc.) on a frame-by-frame basi... |
6 | One of a Kind: Inferring Personality Impressions in Meetings.
- Aran, Gatica-Perez
- 2013
(Show Context)
Citation Context ...features extracted from facial expressions [8], Motion Energy Images [4] and other nonverbal cues based on speaking activity, distance to camera, and looking at the camera while speaking. The work in =-=[9]-=- used similar features as well as head and body activity, and social attention features (attention given by the target subject to the others, attention received from the others) in a small group meeti... |
6 | Annotation and processing of continuous emotional attributes: Challenges and opportunities,” in
- Metallinou, Narayanan
- 2013
(Show Context)
Citation Context ...onality assessment approach based on continuous prediction of perceived trait and social dimensions in space and time. Although continuous prediction of affect has been well studied in the literature =-=[11, 12]-=-, to the best of our knowledge, such changes in the personality impressions have not been explored yet. Annotations obtained continuously in time and in trait-social dimension have the potential to pr... |
5 |
Hi YouTube!: Personality Impressions and Verbal Content in Social Video.
- Biel, Tsiminaki, et al.
- 2013
(Show Context)
Citation Context ...ions, namely, agreeableness, conscientiousness, extroversion, neuroticism and openness to experience. Although most of the existing literature focused on a subset of personality of the Big Five model =-=[1, 2, 3, 4]-=-, few studies also took into account other social dimensions such as likeability [5], group involvement and individual engagement [6], persuasiveness as well as its potential correlation with the five... |
4 |
Multimodal recognition of personality traits in human-computer collaborative tasks,” in
- Batrinca, Lepri, et al.
- 2012
(Show Context)
Citation Context ...ions, namely, agreeableness, conscientiousness, extroversion, neuroticism and openness to experience. Although most of the existing literature focused on a subset of personality of the Big Five model =-=[1, 2, 3, 4]-=-, few studies also took into account other social dimensions such as likeability [5], group involvement and individual engagement [6], persuasiveness as well as its potential correlation with the five... |
4 |
Who is persuasive?: The role of perceived personality and communicationmodality in social multimedia,” in
- Mohammadi, Sagae, et al.
- 2013
(Show Context)
Citation Context ...o took into account other social dimensions such as likeability [5], group involvement and individual engagement [6], persuasiveness as well as its potential correlation with the five main dimensions =-=[7]-=-. To predict five major personality dimensions, Batrinca et al. [1] used short self-presentations where the subjects were asked to introduce themselves in front of a camera. Each subject, at the same ... |
4 |
On the relationship between head pose, social attention and personality prediction for unstructured and dynamic group interactions
- Subramanian, Yan, et al.
- 2013
(Show Context)
Citation Context ...de Regression and SVR classifiers from Youtube video blogs, and then the trained classifiers were tested on small group meeting data for recognizing extroversion. In the same vein, Subramanian et al. =-=[10]-=- also explored features extracted from thin slices, in particular, proximity features such as distance between the target subject and the others, the velocity of the target subject in a given time win... |
3 |
A graphical user interface for continuous annotation of non-verbal signals,”
- Motichande
- 2013
(Show Context)
Citation Context ... length was appeared to be sufficiently long to capture these behavioural changes and was reasonable for obtaining effective annotations. Annotation. For the annotation task, we used an in-house tool =-=[16]-=- that requires the annotator to scroll a bar between a range of values (1 and 100) along time. The clips were annotated by 21 paid participants aged between 23 and 53 years (mean = 29). Each participa... |
3 |
Local Zernike moment representations for facial affect recognition,” in
- Sariyanidi, Gunes, et al.
- 2013
(Show Context)
Citation Context ... automatic continuous prediction problems. For example, in the context of affect recognition, a similar framework obtained the best correlation value of COR= 0.22 for predicting the valence dimension =-=[20]-=-. Overall, the proposed method finds a plausible relationship (COR> 0.1) for all dimensions. Table 3 also indicates that Ridge regression provides better prediction results compared to Lasso regressio... |
3 | Automatic prediction of perceived traits using visual cues under varied situational context,” in
- Joshi, Gunes, et al.
- 2014
(Show Context)
Citation Context ...e/less extravertedly depending on the situation. A recent study on automatic prediction of perceived traits also showed that interactive context affects the trait perception of the external observers =-=[22]-=-. Therefore, investigating the effects of situational context appears to be an interesting research question to focus on as future work. Acknowledgment This work is supported by the EPSRC MAPTRAITS Pr... |
2 |
Cross-domain personality prediction: From blogs to small group meetings,” in
- Aran, Gatica-Perez
- 2013
(Show Context)
Citation Context ...ions, namely, agreeableness, conscientiousness, extroversion, neuroticism and openness to experience. Although most of the existing literature focused on a subset of personality of the Big Five model =-=[1, 2, 3, 4]-=-, few studies also took into account other social dimensions such as likeability [5], group involvement and individual engagement [6], persuasiveness as well as its potential correlation with the five... |
2 | Likability of human voices: A feature analysis and a neural network regression approach to automatic likability estimation,” in
- Eyben, Weninger, et al.
- 2013
(Show Context)
Citation Context ...erience. Although most of the existing literature focused on a subset of personality of the Big Five model [1, 2, 3, 4], few studies also took into account other social dimensions such as likeability =-=[5]-=-, group involvement and individual engagement [6], persuasiveness as well as its potential correlation with the five main dimensions [7]. To predict five major personality dimensions, Batrinca et al. ... |
2 | A gaze-based method for relating group involvement to individual engagement in multimodal multiparty dialogue
- Oertel, Salvi
- 2013
(Show Context)
Citation Context ... focused on a subset of personality of the Big Five model [1, 2, 3, 4], few studies also took into account other social dimensions such as likeability [5], group involvement and individual engagement =-=[6]-=-, persuasiveness as well as its potential correlation with the five main dimensions [7]. To predict five major personality dimensions, Batrinca et al. [1] used short self-presentations where the subje... |
1 |
Tracing emotion: an overview,” Int
- Cowie, McKeown, et al.
- 2012
(Show Context)
Citation Context ...onality assessment approach based on continuous prediction of perceived trait and social dimensions in space and time. Although continuous prediction of affect has been well studied in the literature =-=[11, 12]-=-, to the best of our knowledge, such changes in the personality impressions have not been explored yet. Annotations obtained continuously in time and in trait-social dimension have the potential to pr... |
1 |
effect,” http://en.wikipedia. org/wiki/Halo_effect, accessed at
- Wikipedia
- 2014
(Show Context)
Citation Context ...person in the given context). While the engagement dimension is for evaluating the interaction between a human subject and virtual characters, other social dimensions are related to the “Halo Effect” =-=[13]-=- which states that observers tend to assign good attributes to the person that they find attractive or they like. Measuring the correlation between these dimensions and the Big Five model can offer in... |