Human interaction categorization by using audio-visual cues

Marín-Jiménez, Manuel J.; Muñoz-Salinas, Rafael; Yeguas-Bolivar, Enrique; Blanca, Nicolás Pérez de la

doi:10.1007/s00138-013-0521-1

Cited by 26 publications

(16 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Social networking methods model the human behavior from gestures, body motion and speech (Patron-Perez, A., Marszalek, M., Reid, I., and Zisserman, A. (2012) [18] ; Marn-Jimnez, M. J., Noz Salinas, R. M., Yeguas-Bolivar, E., and de la Blanca, N. P. (2014) [19]). By estimating the orientation and location of the faces of persons and computing a line of sight for each face we can model social interactions and we can get the location of the individuals Fathi, A., Hodgins, J. K., and Rehg, J. M. (2012) [20] and by modeling relationships between interacting person we can predict joint social interaction Park, H. S., and Shi, J.…”

Section: Social Networking Methodsmentioning

confidence: 99%

Human Activity Recognition System (Hars) Using Histogram of Normalized Fourier Descriptor

Samir

El-Munim

Aly

2017

Journal of Al-Azhar University Engineering Sector

View full text Add to dashboard Cite

This paper presents a system, which is able to recognize different continuous human activities in real time from videos using a single stationary camera. The proposed system has a motion descriptor, Called histogram of normalized Fourier descriptor of object contour. After image segmentation the proposed system extract the object contour and estimate the FD "Fourier descriptor" of that contour then normalize FD to cancel the effect of starting point variation ,Rotation and Scale. The histogram of the normalized Fourier descriptor is used as a feature vector.The authors used three methods for classification Support Vector Machine,One versus all support vector machine and Naive Bayes classifier. The classification by using SVM shows better performance than other methods. Experimental results on two data sets weizman and KTH validate the proposed system reliability and efficiency.

show abstract

Section: Social Networking Methodsmentioning

confidence: 99%

Human Activity Recognition System (Hars) Using Histogram of Normalized Fourier Descriptor

Samir

El-Munim

Aly

2017

Journal of Al-Azhar University Engineering Sector

View full text Add to dashboard Cite

show abstract

“…Shao et al [22] mixed appearance and motion features using multi-task deep learning for recognizing group activities in crowded scenes collected from the web. Marín-Jiménez et al [23] used a bag of visualaudio words scheme along with late fusion for recognizing human interactions in TV shows. Even though their method performs well in recognizing human interaction, the lack of an intrinsic audio-visual relationship estimation limits the recognition problem.…”

Section: Related Workmentioning

confidence: 99%

Inferring Human Activities Using Robust Privileged Probabilistic Learning

Vrigkas

Kazakos

Nikou

et al. 2017

2017 IEEE International Conference on Computer Vision Workshops (ICCVW)

View full text Add to dashboard Cite

In this work, a novel method based on the learning using privileged information (LUPI) paradigm for recognizing complex human activities is proposed that handles missing information during testing. We present a supervised probabilistic approach that integrates LUPI into a hidden conditional random field (HCRF) model. The proposed model is called HCRF+ and may be trained using both maximum likelihood and maximum margin approaches. It employs a self-training technique for automatic estimation of the regularization parameters of the objective functions. Moreover, the method provides robustness to outliers (such as noise or missing data) by modeling the conditional distribution of the privileged information by a Student's t-density function, which is naturally integrated into the HCRF+ framework. Different forms of privileged information were investigated. The proposed method was evaluated using four challenging publicly available datasets and the experimental results demonstrate its effectiveness with respect to the-state-of-the-art in the LUPI framework using both hand-crafted features and features extracted from a convolutional neural network.Index Terms-Hidden conditional random fields, learning using privileged information, human activity recognition ✦

show abstract

“…Gaidon et al [16] propose to use clustertrees of tracklets. Marin-Jimenez et al [27] proposes to use audio features, as well as visual features for interaction recognition. While audio features can be useful, in this work, we focus solely on visual features, and show that without the need for complex models, our simple framework of two-person based visual features coupled with spatial multiple instance embedding proves to be an effective way for two-person interaction recognition.…”

Section: Related Workmentioning

confidence: 99%

Two-person interaction recognition via spatial multiple instance embedding

Sener

Ikizler-Cinbis

2015

Journal of Visual Communication and Image Representation

View full text Add to dashboard Cite

a b s t r a c tIn this work, we look into the problem of recognizing two-person interactions in videos. Our method integrates multiple visual features in a weakly supervised manner by utilizing an embedding-based multiple instance learning framework. In our proposed method, first, several visual features that capture the shape and motion of the interacting people are extracted from each detected person region in a video. Then, twoperson visual descriptors are formed. Since the relative spatial locations of interacting people are likely to complement the visual descriptors, we propose to use spatial multiple instance embedding, which implicitly incorporates the distances between people into the multiple instance learning process. Experimental results on two benchmark datasets validate that using two-person visual descriptors together with spatial multiple instance learning offers an effective way for inferring the type of the interaction.

show abstract

Human interaction categorization by using audio-visual cues

Cited by 26 publications

References 26 publications

Human Activity Recognition System (Hars) Using Histogram of Normalized Fourier Descriptor

Human Activity Recognition System (Hars) Using Histogram of Normalized Fourier Descriptor

Inferring Human Activities Using Robust Privileged Probabilistic Learning

Two-person interaction recognition via spatial multiple instance embedding

Contact Info

Product

Resources

About