Automated Facial Feature Detection from Portrait and Range Images

Jahanbin, Sina; Bovik, Alan C.; Choi, Hyun-Jun

doi:10.1109/ssiai.2008.4512276

Cited by 23 publications

(12 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to find the location of facial landmarks on any incoming coregistered range and portrait pairs, we have adapted a multimodal (2D+3D) feature point localization algorithm described in [3]. In the training phase, a set of 50 pairs of registered range and gray scale portrait images was selected in [3] A set of 11 prominent facial features were manually marked only on the portrait images of these 50 training pairs.…”

Section: Automatic Facial Feature Detectionmentioning

confidence: 99%

“…In the training phase, a set of 50 pairs of registered range and gray scale portrait images was selected in [3] A set of 11 prominent facial features were manually marked only on the portrait images of these 50 training pairs. Since these portrait and range images are perfectly aligned, the location of fiducials on the range image of the pair is exactly the same as the portrait one.…”

Section: Automatic Facial Feature Detectionmentioning

confidence: 99%

“…As a result, each pixel of the search area is represented with two Gabor jets, a "range jet" and a "portrait jet". Next, the jets of each modality are compared to their corresponding bunch using the similarity measure between a jet and a bunch formulated by (3). Consequently, a similarity map is created for each modality demonstrating the similarity between each pixel in the search area and the appropriate bunch describing the appearance of the target feature point.…”

Section: Automatic Facial Feature Detectionmentioning

confidence: 99%

“…Inspired by the EBGM idea, we have employed Gabor wavelets to automatically and accurately detect multiple fiducial points on portrait and range images of faces [3]. Each fiducial point was co-localized using both portrait and range Gabor jets.…”

Section: Introductionmentioning

confidence: 99%

“…A predetermined search window was placed on a fiducial, and a jet was computed at every pixel in the window. The location of a jet having the maximum similarity to the jets of training data was identified as the precise location of the particular fiducial (please refer to [3] for details). After the automatic fiducial point localization, Gabor jets of all fiducial points are concatenated in each modality (2D and 3D), creating two independent feature vectors per face.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Automated facial feature detection and face recognition using Gabor features on range and portrait images

Jahanbim

Choi

Jahanbin

et al. 2008

2008 15th IEEE International Conference on Image Processing

Self Cite

View full text Add to dashboard Cite

In this paper, we present a novel identity verification system based on Gabor features extracted from range (3D) representations of faces. Multiple landmarks (fiducials) on a face are automatically detected using these Gabor features. Once the landmarks are identified, the Gabor features on all fiducials of a face are concatenated to form a feature vector for that particular face. Linear discriminant analysis (LDA) is used to reduce the dimensionality of the feature vector while maximizing the discrimination power. These novel features were tested on 1196 range images. The same features were also extracted from portrait images, and the accuracies of both modalities were compared. A superior verification accuracy was obtained using the range data, and a highly competitive accuracy to that of other techniques in the literature was also obtained for the portrait data.

show abstract

Section: Automatic Facial Feature Detectionmentioning

confidence: 99%

Section: Automatic Facial Feature Detectionmentioning

confidence: 99%

Section: Automatic Facial Feature Detectionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations