Datasets
We provide two kinship datasets: KinFaceW-I and KinFaceW-II. Face images are collected from the internet, including some public figure face images as well as their parents' or children's face images. Face images are captured under uncontrolled environments in two datasets with no restriction in terms of pose, lighting, background, expression, age, ethnicity, and partial occlusion. The difference of KinFaceW-I and KinFaceW-II is that face images with a kin relation were acquired from different photos in KinFaceW-I and the same photo in KinFaceW-II in most cases.
There are four kin relations in two datasets: Father-Son (F-S), Father-Daughter (F-D), Mother-Son (M-S), and Mother-Daughter (M-D). In the KinFaceW-I dataset, there are 156, 134, 116, and 127 pairs of kinship images for these four relations. For the KinFaceW-II dataset, each relation contains 250 pairs of kinship images. For ease of use, we manually labeled the coordinates of the eyes position of each face image, and aligned and cropped facial region into 64x64 to remove background. The following figures show some cropped face images from KinFaceW-I and KinFaceW-II datasets, respectively.
Figure 1 & 2. Aligned and cropped examples of KinFaceW-I and KinFaceW-II datasets, respectively. From top to bottom are the F-S, F-D, M-S and M-D kinship relations, and the neighboring two images in each row are with the kinship relation, respectively.
Kinship Face Videos in the Wild (KFVW), a video face dataset, was collected from TV shows on the Web for the video-based kinship verification study. Compared to a still image, a video provides more information to describe the appearance of human face, because it can easily capture the face of the person of interest from different poses, expressions, and illuminations. We totally collected 418 pairs of face videos, and each video contains about 100-500 frames with large variations such as pose, lighting, background, occlusion, expression, makeup, age, etc. The average size of a video frame is about 900 × 500 pixels. There are also four kinship relation types in the KFVW dataset: Father-Son (F-S), Father-Daughter (F-D), Mother-Son (M-S), and Mother-Daughter (M-D), and there are 107, 101, 100, and 110 pairs of kinship face videos for F-S, F-D, M-S, and M-D respectively.
Figure 3. Sampled video frames of our KFVW dataset. Each row lists three face images of a video.