T. Kaya and K. Kobayashi, "A basic study on human face recognition," in Frontiers of Pattern Recognition, S. Watanabe, Ed. New York: Academic, pp. 265--289, 1972.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....(b) a video clip from TV news Figure 1.1: Examples of the video sequences that we are interested 1.3 Objective The problem of face recognition has been investigated for over twenty years and lots of different techniques have been reported. Those techniques include geometric feature matching [19, 20, 22], template matching [22] deformable template matching [26] Karhunen Lo eve Transformation (KLT) 14] singular value decomposition (SVD) isodensity lines [5] neural network [28, 29] etc. Although high recognition rate has been obtained in some of these works, most of the authors ....

....various approaches to extracting features from a face have been proposed. In the following subsections, we will briefly review some of the most well known approaches. 2.3. 1 Geometric Feature Approaches This approach was popular in the early stage of the development of face recognition technology [19, 20, 39, 40]. The idea is trying to measure geometric parameters of a face, such as the size of eyes. the distance between nose to mouth, the size of the nose, the width of the mouth, the shape of the chin, etc. Moreover, the measured geometric features must be somehow normalized in order to be ....

T. Kaya and K. Kobayashi, "A basic study on human face recognition," in Frontiers of Pattern Recognition, S. Watanabe, Ed. New York: Academic, pp. 265--289, 1972.

....features are unknown. This usually leads to a decrease in the dimension of the input data. The hope is that the remaining information is information about local geometry at a very coarse level. The success of this approach depends on accurate extraction of the features. See Goldstein ( 49] Kaya ([71]) Bledsoe ( 25] and Kanade ( 68] for examples of this approach. 6.1.2 Template Matching Based Approach In its original form, template matching proceeds by matching the image (repre sented as a 2 dimensional matrix) to a single template for the whole face using some metric. More sophisticated ....

Y. Kaya and K. Kobayashi, "A Basic Study on Human Face Recognition", in Frontiers of Pattern Recognition (S. Watanabe ed.), p. 265, 1972

....sporadic over the years, evidently due to the difficulty of extracting meaningful information from facial images. Facial classification systems based on measurements derived from interactively selected fiducial points (eye and mouth comers, nose, top of head, etc. go back to the mid 1960 s [2] [15], 11] Early attempts at recognition through automated facial feature identification include [25] and [13] Part of the difficulty of facial image analysis is that the face is highly deformable, particularly around the forehead, eyes, and mouth, and these deformations convey a great deal of ....

Y. Kaya and K. Kobayashi, "A basic study on human face recognition," in Frontiers of Pattern Recognition (S. Watanabe, Ed.). 1972, p. 265.

....filtering scheme based on steerable filters. Keywords: Keypoint and feature detection, knowledge integration, face recognition, steerable filters 1. Introduction Facial keypoints such as eye corners and mouth ends are important features for many different tasks in automatic face processing [4, 5]. The localization of facial keypoints is usually performed interactively or it is not very precise and robust. In general, the problem is that anatomical facial landmarks we are searching for in this paper are defined rather as morphological features (e.g. the corner of an eye) than by a ....

Y. Kaya et al., A basic study on human face recognition, in S. Watanabe (Ed.), Frontiers of Pattern Rec., Academic Press, 265-289, 1972.

....extremities, nostrils and chin top. The features defined for face profiles (side views) typically include a set of characteristic points on the profile (such as the notch between the brow and the nose or the tip of the nose) and the angles between these points. For example, Kaya and Kobayashi [Kaya and Kobayashi, 1972] used Euclidean distances between manually identified points in the images to characterize the faces. Kanade [Kanade, 1977] used the distances and angles between eye corners, ends of the mouth, nostrils, and top of the chin, but the location of those facial features were extracted automatically by ....

Kaya, Y. and Kobayashi, K. (1972). A basic study on human face recognition. In Wantanabe, S., editor, Frontiers of Pattern Recognition, pages 265--289. Academic Press, New York.

....of tasks typical of human performance. Traditionally, computational models of face recognition represent faces in terms of geometric descriptors that include distances, angles, and areas between elementary features such as eyes, nose, or chin (Harmon Hunt, 1977; Harmon, Khan, Lash Ramig, 1981; Kaya Kobayashi, 1972; Sakai, Nagao Kidode, 1971) or in terms of template parameters (Yuille, 1991) or isodensity lines (Nakamura, Mathur Minami, 1991) Although this approach 1 constitutes an economical way of representing faces that is relatively insensitive to variations in scale, tilt, or rotation of the ....

Kaya, Y. & Kobayashi, K. (1972). A basic study on human face recognition. International Conference on Frontiers of Pattern Recognition, 265--289.

....major aspects of face study: representation of faces, detection of faces, identification of faces, analysis of facial expressions, and classification of faces based on physical features. 8.2. 1 Earlier Systems The early systems for face recognition were essentially manual systems such as Photofit [42, 29, 12, 44]. Alternatives to Photofit system include Multiple Image Maker and Identification Compositor (MIMIC) Figure 6: Sketch Pad Window for Specifying RSC Queries 38 (uses film strip projections) Identikit (uses plastic overlays of drawn features) and Compusketch (a computerized version of ....

....computer vision researchers. Major focus has been on automatically extracting facial features from the 2D intensity images. The success of these systems is quite limited despite several strong assumptions about the environment in which the image is produced as well as about the pose of the face [29, 12, 44]. Little attention has been paid to the database issues such as feature organization and matching algorithms. In most of the cases, matching algorithm is simply the Euclidean distance between the corresponding features. However, recent research indicates that features stored in terms of their ....

Kaya, Y. and Kobayashi, K. (1987), "A Basic Study on Human Face Recognition," in Frontiers of Pattern Recognition, Academic Press, pp. 265-289.

....accurate knowledge of population statistics of stored face descriptions. These 21 face descriptions are not metric measures but statistics described by integer values, from 1 to 5, as listed in Figure 1.1. Kaya et al. conducted a similar experiment to explore the characteristic parameters of faces [14, 15]. They had ten to forty men look at photographs from three men for a few seconds and report their impressions on characteristics of faces in photographs, and accordingly 9 characteristic parameters were selected as the geometric measurements to the faces. These parameters are: internal bi ocular ....

Y. Kaya and K. Kobayashi. A basic study on human face recognition. In Frontiers of Pattern Recognition, pages 256 -- 289. Academic Press, 1972.

....accurate knowledge of population statistics of stored face descriptions. These 21 face descriptions are not metric measures but statistics described by integer values, from 1 to 5, as listed in Figure 1.1. Kaya et al. conducted a similar experiment to explore the characteristic parameters of faces [14, 15]. They had ten to forty men look at photographs from three men for a few seconds and report their impressions on characteristics of faces in photographs, and accordingly 9 characteristic parameters were selected as the geometric measurements to the faces. These parameters are: internal bi ocular ....

Y. Kaya and K. Kobayishi. A basic study on human face recognition. In Proc. Int. Conf. on Pattern Recognition, pages 265--289, Hawaii, January 1971. 107

....as eyes, nose and mouth) are no longer resolved. The remaining information is, in a sense, purely geometrical and represents what is left at a very coarse resolution. The idea is to extract relative position and other parameters of distinctive features such as eyes, mouth, nose and chin [10] [14], 8] 2] 13] This was the first approach towards an automated recognition of faces [13] Template matching. In the simplest version of template matching, visual patterns, represented as bidimensional arrays of intensity values, are compared using a suitable metric (typically the euclidean ....

....discrimination. The overall configuration can be described by a vector of numerical data representing the position and size of the main facial features: eyes and eyebrows, nose and mouth. This information can be supplemented by the shape of the face outline. As put forward by Kaya and Kobayashi [14] the set of features should satisfy the following requisites: ffl estimation must be as easy as possible; ffl dependency on light conditions must be as small as possible; ffl dependency on small changes of face expression must be small; ffl information contents must be as high as possible. ....

Y. Kaya and K. Kobayashi. A basic study on human face recognition. In S. Watanabe, editor, Frontiers of Pattern Recognition, page 265. 1972.

....in particular the problem of automatic face recognition. This problem has been considered to be a challenge since the very first days of computer vision. One of the first approaches to this problem was based on geometric features, such as size and relative positions of eyes, mouth, nose and chin [3 6]. Another basic technique is template matching which has reached a considerable level of sophistication [7 9] Further approaches to face recognition use graph matching [10] Karhunen Loewe expansion [11,12] algebraic moments [13] isodensity lines[14] etc. Connectionists approaches to the ....

Y. Kaya and K. Kobayashi: "A basic study on human face recognition", in S. Watanabe (ed.) Frontiers of Pattern Recognition (1972) pp. 265.

....Face recognition, a problem that has been considered to be a challenge since the very first days of computer vision, recently experiences a revival. One of the first approaches to this problem was based on geometric features, such as size and relative positions of eyes, mouth, nose and chin [7 8]. Another basic technique which has reached a considerable level of sophistication is template matching [9 11] Further approaches to face recognition use graph matching [11 12] Karhunen Loewe expansion [13] algebraic moments [14] isodensity lines[15] etc. Connectionists approaches to the ....

Y. Kaya and K. Kobayashi: "A basic study on human face recognition", in S. Watanabe (ed.) Frontiers of Pattern Recognition (1972) pp. 265.

....in particular the problem of automatic face recognition. This problem has been considered to be a challenge since the very first days of computer vision. One of the first approaches to this problem was based on geometric features, such as size and relative positions of eyes, mouth, nose and chin [3 6]. Another basic technique is template matching which has reached a considerable level of sophistication [7 9,30] Further approaches to face recognition use graph matching [10] Karhunen Loewe expansion [11,12] algebraic moments [13] isodensity lines[14] etc. Connectionists approaches to the ....

Y. Kaya and K. Kobayashi: "A basic study on human face recognition", in S. Watanabe (ed.) Frontiers of Pattern Recognition (1972) pp. 265.

....space. Face recognition, a problem that has been considered to be a challenge since the very first days of computer vision, recently experiences a revival. One of the first approaches to this problem was based on geometric features, such as size and relative positions of eyes, mouth, nose and chin [6, 7, 8]. Other basic techniques, which have reached a considerable level of sophistication, are template and graph matching [9, 10, 11, 12] Further approaches to face recognition use Karhunen Loewe expansion [13] algebraic moments [14] isodensity lines [15] etc. Connectionists approaches to the ....

Y. Kaya and K. Kobayashi: "A basic study on human face recognition", in S. Watanabe (ed.) Frontiers of Pattern Recognition (1972) pp. 265.

....representation, a geometrical approach that uses the spatial configuration of facial features, and a more pictorial approach that uses an image based representation. There have been several feature geometry approaches, beginning with the seminal work of Kanade[21] and including Kaya and Kobayashi[22], Craw and Cameron[13] Wong, Law, and Tsang[41] Brunelli and Poggio[7] and Chen and Huang[10] These feature based systems begin by locating a set of facial features, including such features as the corners of the eyes and mouth, sides of the face and nose, nostrils, the contour along the chin, ....

Y. Kaya and K. Kobayashi. A basic study on human face recognition. In Satosi Watanabe, editor, Frontiers of Pattern Recognition, pages 265--289. Academic Press, New York, NY, 1972.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC