【正文】 184windows to down to 1 in 41。 the receptive fields of hidden units are shown in Fig. 1. There are three types of hidden units: 4 which look at 10x10 pixel subregions, 16 which look at 5x5 pixel subregions, and 6 which look at overlapping 20x5 pixel horizontal stripes of pixels. Each of these types was chosen to allow the hidden units to detect local features that might be important forface detection. In particular, the horizontal stripes allow the hidden units to detect such features as mouths or pairs of eyes, while the hidden units with square receptive fields might detect features such as individual eyes, the nose, or corners of the mouth. Although the figure shows a single hidden unit for each subregion of the input, these units can be replicated. For the experiments which are described later, we use networks with two and three sets of these hidden units. Similar input connection patterns are monly used in speech and character recognition tasks [10, 24]. The network has a single, realvalued output, which indicates whether or not the window contains a face.Examples of output from a single network are shown in Fig. 3. In the figure, each box representsthe position and size of a window to which the neural network gave a positive response. The network has some invariance to position and scale, which results in multiple boxes around some faces. Note also that there are some false detections。613。Rowley, Baluja, and Kanade: Neural NetworkBased Face Detection (PAMI, January 1998)24Copyright 1998 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional pur poses or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted ponent of this work in other works must be obtained from the IEEE.Neural NetworkBased Face DetectionHenry A. Rowley, Shumeet Baluja, and Takeo Kanade AbstractWe present a neural networkbased upright frontal face detection system. A retinally con nected neural network examines small windows of an image, and decides whether each win dow contains a face. The system arbitrates between multiple networks to improve performance over a single network. We present a straightforward procedure for aligning positive face ex amples for training. To collect negative examples, we use a bootstrap algorithm, which adds false detections into the training set as training progresses. This eliminates the difficult task of manually selecting nonface training examples, which must be chosen to span the entire space of nonface images. Simple heuristics, such as using the fact that faces rarely overlap in images, can further improve the accuracy. Comparisons with several other stateoftheart face detec tion systems are presented。605, depending on the type of arbitration used. Systems 10, 11, and 12 show that the detector can be tuned to make it more or less conservative. System 10, which uses ANDing, gives an extremely small number of false positives, and has a detection rate of about %. On the other hand, System 12, which is based on ORing, has a higher detection rate of % but also has a larger number of false detections. System 11 provides a promise between the two. The differences in performance of these systems can be understood by considering the arbitration strategy. When using ANDing, a false detection made by only one network is suppressed, leading to a lower false positive rate. On the other hand, when ORing is used, faces detected correctly by only one network will be preserved, improving the detection rate.Systems 14, 15, and 16, all of which use neural networkbased arbitration among three networks, yield detection and false alarm rates between those of Systems 10 and 11. System 13, which uses voting among three networks, has an accuracy between that of Systems 11 and 12. System 17 will be described in the next section. Table 2 shows the result of applying each of the systems to images in Test Set 2 (a subset of public portion of the FERET database [16, 17]). We partitioned the images into three groups, based on the nominal angle of the face with respect to the camera: frontal faces, faces at an angle 15 4from the camera, and faces at an angle of 22:5. The direction of the face varies significantlywithin these groups. As can be seen from the table, the detection rate for systems arbitrating two networks ranges between % and % for frontal and 15faces, while for 22:5faces, thedetection rate is between % and %. This difference is because the training set contains mostly frontal faces. It is interesting to note that the systems generally have a higher detection rate for faces at an angle of 15than for frontal faces. The majority of people whose frontal faces aremissed are wearing glasses which are reflecting light into the camera. The detector is not trained on such images, and expects the eyes to be darker than the rest of the face. Thus the detection rate for such faces is lower.Based on the results shown in Tables 1 and 2, we concluded that both Systems 11 and 15make acceptable tradeoffs between the number of false detections and the detection rate. Because System 11 is less plex than System 15 (using only two networks rather than a total of four), it is preferable. System 11 detects on average % of the faces, with an average of one false detection per 3。 they will be eliminated by methods presented in Section .TTo train the neural network used in stage one to serve as an accurate filter, a large number offace and nonface images are needed. Nearly 1050 face