Probabilistic Visual Learning for Object Representation
IEEE Transactions on Pattern Analysis and Machine Intelligence
Neural Network-Based Face Detection
IEEE Transactions on Pattern Analysis and Machine Intelligence
Example-Based Learning for View-Based Human Face Detection
IEEE Transactions on Pattern Analysis and Machine Intelligence
Training Support Vector Machines: an Application to Face Detection
CVPR '97 Proceedings of the 1997 Conference on Computer Vision and Pattern Recognition (CVPR '97)
Probabilistic Modeling of Local Appearance and Spatial Relationships for Object Recognition
CVPR '98 Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
Face detection using quantized skin color regions merging andwavelet packet analysis
IEEE Transactions on Multimedia
Detecting Facial Features on Image Sequences Using Cross-Verification Mechanism
PCM '01 Proceedings of the Second IEEE Pacific Rim Conference on Multimedia: Advances in Multimedia Information Processing
Face Indexing and Retrieval in Personal Digital Album
PCM '01 Proceedings of the Second IEEE Pacific Rim Conference on Multimedia: Advances in Multimedia Information Processing
| Hi-index | 0.00 |
In this paper a general framework for face detection is presented which taken into accounts both color and gray level images. For color images, skin color segmentation is used as the first stage to reduce search space into a few gray level regions possibly containing faces. And then in general for gray level images, techniques of template matching based on average face for searching face candidates and neural network classification for face verification are integrated for face detection. A qualitative 3D model of skin color in HSI space is used for skin color segmentation. Two types of templates: eyes-in-whole and face itself, are used one by one in template matching for searching face candidates. Two three-layer-perceptrons (MLPs) are used independently in the template matching procedure to verify each face candidate to produce two list of detected faces which are arbitrated to exclude most of the false alarms. Experiment results demonstrate the feasibility of this approach.