The Daxing dataset includes 43,400 images and 1,400 videos captured by 90 smartphones of 22 models belonging to five brands. For example, there are 23 smartphone devices for iPhone 6S (Plus) model. To the best of our knowledge, the Daxing dataset uses the largest amount of smartphones for image/video source identification compared with other related datasets, as well as the highest numbers of devices per model and captured images/videos. The dataset will be shared as an open source for scientific researchers and criminal investigators.
All the images in this dataset were directly copy from smartphones without postprocessing. The resolution of the images is the default resolution of each smartphone camera. The resolution of smartphone is as follows. For the convenience, we have encoded every device. In the encoding process, the first code of all smartphones of the same brand is the same, the first two codes of all smartphones of the same model are the same, and the last two codes represent different individual devices of the same model. The specific codes are
shown as follows.
Fetch Code: 2024
We selected the popular smartphone brands on the market currently for image collection, including Huawei, Apple, OPPO, VIVO, Xiaomi. Each brand has at least 5 smartphone devices.
After the devices were selected, the image collection work needed to be carried out. The image collection work mainly includes the selection of the shooting scene, the photographing setting of the imaging device, and the number of images. Selected scenes include “sky”, “grass”, “stone”, “trees”, “staircase”, “indoor vertical printer”, “lobby wall”, “white wall in the classroom”, etc. In each scene, imageswere captured when smartphone was placed at three different angles (we regarded the vertical position of the smartphone as the reference, and the placement angle was set as 90 degrees, 0 degrees when rotating 90 degrees counterclockwise and 180 degrees when rotating 90 degrees clockwise, respectively). All of cameras were set to “Default” mode and flash was set to “Off” mode in the collection process. Specific information including scenes display angles and image numbers is shown in Table II. The sky scene provides the most amount of images, no less than 102 per smartphone. For the other seven scenes, there are no less than 50 photos of each scene. As a result, each smartphone captures no less than 482 images in eight scenes. Some examples in the Daxing dataset are shown as follows.
H. Tian, Y. Xiao, G. Cao, Y. Zhang, Z. Xu and Y. Zhao, "Daxing Smartphone Identification Dataset," in IEEE Access, vol. 7, pp. 101046-101053, 2019.
DOI: 10.1109/ACCESS.2019.2928356
URL: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8760241&isnumber=8600701