Multiview structure from motion in trajectory space

Zaheer, Aamer; Akhter, Ijaz; Baig, Mohammad Haris; Marzban, Shabbir; Khan, Sohaib

doi:10.1109/iccv.2011.6126529

Cited by 12 publications

(18 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…These additional constraints above were called basis constraints [30][31][32][33][34]. They removed the inherent basis ambiguity by doing this.…”

Section: Shape Basismentioning

confidence: 99%

“…A large number of efforts had been made for it, and the result is delightful. For example, Aamer Zaheer et al [34] made a big contribution to provide algorithms for recovering the non-rigid structure from motion well with missing data by using DCT as well as Principle Component Analysis (PCA) basis. A large number of experiments showed that trajectory had obvious advantages in handling with occlusion data for the smoothness property of curves.…”

Section: Trajectory Basismentioning

confidence: 99%

See 1 more Smart Citation

Research on Non-rigid Structure from Motion: A Literature Review

Yan¹

2015

JFBI

View full text Add to dashboard Cite

Non-rigid Structure from Motion (NRSfM) is a classical computer vision problem. And the main methods used to solve it are in general based on shape models or trajectory models. This paper will provide an overview over kinds of solutions proposed in these researches. It not only gives out the theoretical insights proposed by researchers in recent years, but also discusses them with their pros and cons. At the same time, the progress of the research about this topic is described in detail and its long-term trend is introduced at the end. This paper is very easy to understand, which mainly introduces two practical, everyday models for the NRSfM problem, namely trajectories based model and shape based model. Both of them are based on matrix factorization technology. Inevitably, some relevant optimization methods are mentioned to solve the projection matrix and corresponding coefficients effectively.

show abstract

“…These additional constraints above were called basis constraints [30][31][32][33][34]. They removed the inherent basis ambiguity by doing this.…”

Section: Shape Basismentioning

confidence: 99%

Section: Trajectory Basismentioning

confidence: 99%

Research on Non-rigid Structure from Motion: A Literature Review

Yan¹

2015

JFBI

View full text Add to dashboard Cite

show abstract

“…Temel bileşen vektörleri elde edildikten sonra her bir çerçeve için öznitelik vektörlerinin hesaplanma aşamasına (5) geçiliyor. Bu aşamada çerçeve içinde boş değer ile karşılaştığımızda ise bu değere sahip işaretçi( kolonu) çıkartılıyor ve temel bileşen vektörlerinde o işaretçiye karşılık gelen kısım( satırı) da çıkartılarak öznitelik vektör değerlerinin hala hesaplanabilir durumda kalması sağlanıyor [17]. Bu sayede verideki yerel kayıplar TBA uygulanırken telafi edilmiş oluyor.…”

Section: Deneylerunclassified

Affect burst recognition using multi-modal cues

Turker

Marzban

Erzin

et al. 2014

2014 22nd Signal Processing and Communications Applications Conference (SIU)

Self Cite

View full text Add to dashboard Cite

ÖZETÇEDuygulanımların konuşma sırasında sözel olmayan dışavurumları olarak tanımlanabilen duygusal patlamalar, duygusal durum analizi için önemli ipuçları taşır. Literatürde yalnızca ses bilgisi kullanarak duygusal patlama sezim ve tanıma konularını araştıran çalışmalar mevcut olsa da, çokkipli bir sistem ile ipuçlarını birleştirme üzerine şimdiye dek çok az çaba sarfedilmiştir. Yüz devinimi duygusal patlamaları karakterize eden önemli ipuçlarından biridir; dolayısıyla bu bilginin daha gürbüz bir duygusal patlama sezim ve tanıma amacına yönelik olarak değerlendirilebileceğini düşünüyoruz. Bu çalışmada veri güdümlü bir yaklaşımla, duygusal patlamalar Saklı Markov Modelleri (SMM) ile betimlenmekte ve duygusal patlamaların sınıflandırılması için ses ve yüz deviniminden elde edilen ipuçlarını birleştiren çok kipli bir karar tümleştirme yöntemi kullanılmaktadır. Yöntemimizi sınamak için kullandığımız işitsel-görsel veritabanı çeşitli ikili konuşmalara ait ses ve yüz devinim bilgilerini içermektedir. Bu veritabanı üzerinde gerçekleştirdiğimiz deneyler yüz devinim bilgisinin duygusal patlama tanıma başarımını kayda değer ölçüde arttırdığını göstermektedir. Anahtar Kelimeler: duygusal patlama, çok-kipli tanıma ABSTRACTAffect bursts, which are nonverbal expressions of emotions in conversations, play a critical role in analyzing affective states. Although there exist a number of methods on affect burst detection and recognition using only audio information, little effort has been spent for combining cues in a multimodal setup. We suggest that facial gestures constitute a key component to characterize affect bursts, and hence have potential for more robust affect burst detection and recognition. We take a data-driven approach to characterize affect bursts using Hidden Markov Models (HMM), and employ a multimodal decision fusion scheme that combines cues from audio and facial gestures for classification of affect bursts. We demonstrate the contribution of facial gestures to affect burst recognition by conducting experiments on an audiovisual database which comprise speech and facial motion data belonging to various dyadic conversations.

show abstract

“…The methods addressing this problem can be divided into three categories: imputation, alternation and non-linear optimisation. Imputation algorithms attempt to fill in the missing data entries using complete subset of the data [18,22]. These methods are simple but cannot handle real data, which often tend to be very noisy.…”

Section: Introductionmentioning

confidence: 99%

3D Deformable Shape Reconstruction with Diffusion Maps

Tao¹,

Matuszewski²

2013

Procedings of the British Machine Vision Conference 2013

View full text Add to dashboard Cite

This paper presents a method for recovering deformable shape and motion from uncalibrated 2D video sequence in the presence of missing data. Highly deformable shapes are hard to describe under previously used assumptions, such as global constraint enforcing shapes to lie within a linear subspace. Considering that the data dimensionality may not represent the true complexity of the problem, we suggest that the shapes can be well-modelled in a low dimensional manifold. However, building a dense representation of the manifold requires a large amount of training data which is not feasible in many real applications. The main contribution of this paper is to propose a novel approach for estimating accurate 3D reconstructions utilising manifold learned from a relatively small number of training samples. The problem is addressed by grouping shapes into evolving clusters, with the shapes in each cluster represented in the linear subspace, estimated based on the observations and the prior learned manifold. Results are presented using motion capture data and real video sequences, showing that the proposed method can better model shapes with complex deformations compare to several state-of-the-art techniques, and is robust against noise and missing data.

show abstract

Multiview structure from motion in trajectory space

Cited by 12 publications

References 23 publications

Research on Non-rigid Structure from Motion: A Literature Review

Research on Non-rigid Structure from Motion: A Literature Review

Affect burst recognition using multi-modal cues

3D Deformable Shape Reconstruction with Diffusion Maps

Contact Info

Product

Resources

About