Face alignment robust to occlusion

Roh, Myung-Cheol; Oguri, Takaharu; Kanade, Takeo

doi:10.1109/fg.2011.5771404

Cited by 13 publications

(13 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The experiments are designed to demonstrate the validity of our active adjustment method, illustrate the intuition behind the validation framework, and evaluate the quantitative and qualitative performance of our refinement approach as a whole. Face Alignment For our experiments, we use a face alignment method, based upon [13,3], that is robust to occlusions and approximates face shape S and returns N binary labels corresponding to the estimated state of occlusion for each individual landmark point in S. Since our approach was not designed to be robust for occluded landmark points, during refinement we limit our adjustments to the non-occluded points in order to minimize the number of misalignment cases attributed to occlusions. Datasets We demonstrate the efficacy of our contour refinement approach, by evaluating it on three different face datasets namely, HELEN [11], "Labeled Face Parts in the Wild" (LFPW) [2] and "Annotated Faces in the Wild" .…”

Section: Methodsmentioning

confidence: 99%

“…Active Shape Models [6] and Active Appearance Models [5] are the most well known and widely used models for shape-fitting. Constrained Local Models [13,17,7] are another class of approaches for face alignment that are largely focused on global spatial models built on top of local landmark detectors. Recently many discriminative shapebased regression approaches [4,16] have been proposed in the literature.…”

Section: Related Workmentioning

confidence: 99%

“…Based on observations over various state-of-the-art face alignment results [4,12,15,13], we propose a two-step approach for fixing facial contour misalignment. In the first step, we introduce the active adjustment algorithm responsible for shifting individual landmark points that constitute the facial contour.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Face Alignment Refinement

Zeng

Boddeti

Kanade

2015

2015 IEEE Winter Conference on Applications of Computer Vision

Self Cite

View full text Add to dashboard Cite

Achieving sub-pixel accuracy with face alignment algorithms is a difficult task given the diversity of appearance in real world facial profiles. To capture variations in perspective, occlusion, and illumination with adequate precision, current face alignment approaches rely on detecting facial landmarks and iteratively adjusting deformable models that encode prior knowledge of facial structure. However, these methods involve optimization in latent sub-spaces, where user-specific face shape information is easily lost after dimensionality reduction. Attempting to retain this information to capture this wide range of variation requires a large training distribution, which is difficult to obtain without high computational complexity. Subsequently, many face alignment methods lack the pixel-level accuracy necessary to satisfy the aesthetic requirements of tasks such as face deidentification, face swapping, and face modeling. In many such applications, the primary source of aesthetic inadequacy is a misaligned jawline or facial contour. In this work, we explore the idea of an image-based refinement method to fix the landmark points of a misaligned facial contour. We propose an efficient two stage process -an intuitively constructed edge detection based algorithm to actively adjust facial contour landmark points, and a datadriven validation system to filter out erroneous adjustments. Experimental results show that state-of-the-art face alignment combined with our proposed post-processing method yields improved overall performance over multiple face image datasets.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Face Alignment Refinement

Zeng

Boddeti

Kanade

2015

2015 IEEE Winter Conference on Applications of Computer Vision

Self Cite

View full text Add to dashboard Cite

show abstract

“…To overcome this shortcoming, some researchers have proposed methods to cope with occlusion handling. For example, Roh et al [19] used a large amount of facial feature detectors to provide over-sufficient landmark candidates and a RANSAC-based hypothesis and test method to robustly determine the whole shape. This method relies heavily on the facial feature detectors and is consequently computationally demanding.…”

Section: Related Workmentioning

confidence: 99%

Consensus of Regression for Occlusion-Robust Facial Feature Localization

Lin

Brandt

et al. 2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. We address the problem of robust facial feature localization in the presence of occlusions, which remains a lingering problem in facial analysis despite intensive long-term studies. Recently, regression-based approaches to localization have produced accurate results in many cases, yet are still subject to significant error when portions of the face are occluded. To overcome this weakness, we propose an occlusion-robust regression method by forming a consensus from estimates arising from a set of occlusion-specific regressors. That is, each regressor is trained to estimate facial feature locations under the precondition that a particular pre-defined region of the face is occluded. The predictions from each regressor are robustly merged using a Bayesian model that models each regressor's prediction correctness likelihood based on local appearance and consistency with other regressors with overlapping occlusion regions. After localization, the occlusion state for each landmark point is estimated using a Gaussian MRF semi-supervised learning method. Experiments on both non-occluded and occluded face databases demonstrate that our approach achieves consistently better results over state-of-the-art methods for facial landmark localization and occlusion detection.

show abstract

“…Actually, common face detectors do not perform well under these challenging conditions, which makes face alignment phase more necessary. Recently, Roh et al [39] proposed an occlusion-robust face alignment approach based on a set of feature detectors and the random sample consensus (RANSAC) strategy [13]. In the researches of [44,46], gradient correlation coefficient, an efficient feature for face analysis under occlusions and illumination changes, was employed as a performance criterion for image-to-image alignment.…”

Section: Introductionmentioning

confidence: 99%

Unsupervised joint face alignment with gradient correlation coefficient

Caplier

2015

Pattern Anal Applic

View full text Add to dashboard Cite

International audienceThis work proposes an unsupervised jointalignment framework, referred to as ‘‘Gradient CorrelationCongealing,’’ which aligns an image ensemble bymaximizing a sum of gradient correlation coefficientfunction defined over all images. We, respectively, developtwo different formulations to optimize the objective functionregarding the role of ‘‘template.’’ While most existingface alignment methods suffer from outliers, e.g., occlusions,the proposed algorithms are able to align faces undergoingpartial occlusions. Moreover, our algorithms cancope with nonuniform illumination changes (even extremelydifficult ones), and also, they do not require anypredefined templates. We test the novel approaches againstfour typical joint alignment methods including Least-Squares Congealing, Learned-Miller Congealing, Lucas–Kanade entropy Congealing, and RASL using three challengingface databases: AR, Yale B, and LFW. Experimentalresults prove the efficiency of our approachesunder different conditions, especially when faces are partiallyoccluded, and the proposed algorithms perform muchbetter than all considered method

show abstract

Face alignment robust to occlusion

Cited by 13 publications

References 13 publications

Face Alignment Refinement

Face Alignment Refinement

Consensus of Regression for Occlusion-Robust Facial Feature Localization

Unsupervised joint face alignment with gradient correlation coefficient

Contact Info

Product

Resources

About