Recognizing partial facial action units based on 3D dynamic range data for facial expression recognition

Sun, Yi; Reale, Michael; Liu, Yin

doi:10.1109/afgr.2008.4813336

Cited by 37 publications

(44 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For instance, AAM is very popular in facial expression analysis field and also have been employed for 3D to track faces over their texture images [27,20]. Employing AAMs might be problematic in our framework since we aim to work in a personindependent manner.…”

Section: Action Unit Detection In 3d Based On Nonrigid Registrationmentioning

confidence: 99%

“…This type of 3D registration can also be computationally very expensive. Finally, AAMs was used to track feature points on 3D video [20], though the latter technique requires manually intensive steps for AAM training, needs texture images and is subject-dependent. These techniques are not fully automatic, and they leave a lot of room for improving detailed correspondence work required for a wider range of expressions and/or for detecting subtle expressions.…”

Section: Action Unit Detection In 3d Based On Nonrigid Registrationmentioning

confidence: 99%

“…There has been recently also some work on the recognition of the six universal emotions using 3D video [20]. In this study, first 83 facial features are tracked with subjectdependent Active Appearance Models (AAMs) over projected texture, then Hidden Markov Models are employed to recognize eight AUs of five upper and three lower AUs.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Automatic detection of facial actions from 3D data

Savran

Sankur

2009

2009 IEEE 12th International Conference on Computer Vision Workshops, ICCV Workshops

View full text Add to dashboard Cite

We address the person-independent recognition problem of facial expressions using static 3D face data. The novel approach to the facial expression recognition uses non-rigid registration of surface curvature features. 3D face data is cast onto 2D feature images, which are then subjected to elastic deformations in their parametric space. Each Action Unit (AU) detector is trained over its respective influence domain on the face. The registration task is incorporated in the multiresolution elastic deformation scheme, which yields adequate registration accuracy for mild pose variations. The algorithm is fully automatic and is free of the burden of first localizing anatomical facial points. The algorithm was tested on 22 facial action units of Facial Action Coding System. Promising results obtained indicate that we have an operative device for facial action unit detection, and an intermediate step to infer emotional or mental states. Moreover, experiments conducted with low intensity AU12 -Lip Corner Puller points to the potential of 3D data and the proposed method in subtle expression detection.

show abstract

Section: Action Unit Detection In 3d Based On Nonrigid Registrationmentioning

confidence: 99%

Section: Action Unit Detection In 3d Based On Nonrigid Registrationmentioning

confidence: 99%

See 1 more Smart Citation

Automatic detection of facial actions from 3D data

Savran

Sankur

2009

2009 IEEE 12th International Conference on Computer Vision Workshops, ICCV Workshops

View full text Add to dashboard Cite

show abstract

“…Much of the art in machine learning lies in the selection of features that represent the original data informatively and discriminatively, and a classifier that uses these features to build mathematical models to separate examples into the desired groups. Sun et al [14] used an active appearance model to track 83 predefined positions and then a set of Hidden Markov Models to classify the displacements of these points across a video of 6 frames. Zhao et al [17] used a patch-based 'Statistical Facial feAture Model'(SFAM) to classify features based on three types of data: landmark configuration, local texture, and local geometry.…”

Section: Related Workmentioning

confidence: 99%

Improving 3D Facial Action Unit Detection with Intrinsic Normalization

Yudin¹,

Wetzler²,

Sela³

et al. 2015

Procedings of the Proceedings of the 1st International Workshop on DIFFerential Geometry in Computer Vision for Analysis of Sha

View full text Add to dashboard Cite

Data normalization techniques are commonly used to reduce intra-class variance while preserving inter-class differences that aid in classification and regression tasks. Such practices bring all data examples to similar scale and range, and help to decrease the dimensionality of categorization problems. In this paper we describe a novel use of a geometric framework to extend the concept of data normalization to the domain of functions that lie on surfaces. In this context, normalization is posed as an embedding of all example functions into Riemannian manifolds nearly isometric to one another. Using geometric tools, we propose an implementation for the case of discretized functions on triangulated meshes. To demonstrate the proposed framework we apply it as a preprocessing step to the task of automatic facial action unit detection for depth maps from a 3D scanner. We show how this subject-independent representation produces more accurate action unit classification results. We also empirically demonstrate that utilizing this intrinsic normalization technique indeed improves action unit recognition performance.

show abstract

“…Gesture recognition from digital images and videos is an important topic in computer vision that has been intensively studied over the years. Different methods have been developed for recognizing static [8,58] or dynamic [34] hand gestures, facial expressions [53] or body actions [40].…”

Section: Overview Of Vision-based Gesture Recognitionmentioning

confidence: 99%

Hand landmarks detection and localization in color images

Grzejszczak

Kawulok

Gałuszka

2015

Multimed Tools Appl

View full text Add to dashboard Cite

This paper introduces a new method for detecting and localizing hand landmarks in 2D color images. Location of the hand landmarks is an important source of information for recognizing hand gestures, effectively exploited in a number of recent methods which operate from the depth maps. However, this problem has not yet been satisfactorily solved for 2D color images. Here, we propose to analyze the skin-presence masks, as well as the directional image of a hand using the distance transform and template matching. This makes it possible to detect the landmarks located both at the contour and inside the hand masks. Moreover, we performed an extensive experimental study to compare the proposed method with a number of state-of-the-art algorithms. The obtained quantitative and qualitative results clearly indicate that our approach outperforms other methods, which may help improve the existing gesture recognition systems.

show abstract

Recognizing partial facial action units based on 3D dynamic range data for facial expression recognition

Cited by 37 publications

References 34 publications

Automatic detection of facial actions from 3D data

Automatic detection of facial actions from 3D data

Improving 3D Facial Action Unit Detection with Intrinsic Normalization

Hand landmarks detection and localization in color images

Contact Info

Product

Resources

About