CUHK-CQRC: A Large-Scale Color QR Code Dataset
Examples
Figure 1: Each column presents 2 samples taken under 5 lighting conditions, "fluorescent", "incandescent", "indoor", "outdoor", "shadow".
Description
CUHK-CQRC is a challenging color QR code dataset which consists of 1,506 photos and 3,884 camera previews (video frames) of high-density 3-layer color QR codes captured by different phone models under different lighting conditions. Fig. 1 presents some samples of CUHK-CQRC.
We implement the color QR code generator based on an open-source barcode processing library, ZXing, and we also deploy the HiQ generator on-line at https://authpaper.net/. For fair comparison between HiQ-C and PCCC in which the color QR code is inherently 3-layer, so we generate 5 high-capacity 3-layer color QR codes with different data capacities (excluding redundancies from error correction mechanism) which are 2787 bytes, 3819 bytes, 5196 bytes, 6909 bytes and 8859 bytes (maximum for a 3-layer color QR code). In order to test the limit of each approach, all color QR codes are embedded with low level of error correction in each layer. By using a common color printer (Ricoh Aficio MP C5501A), we print each generated color QR code on ordinary white paper substrate in different printout sizes, $30\times 30\,mm^2$, $40\times 40\,mm^2$, $50\times 50\,mm^2$ and $60\times 60\,mm^2$ (for simplicity, we use the length of one side of the square to represent the printout size), and two different printout resolutions, 600dpi and 1200dpi. To simulate the normal scanning scenario, the samples are captured by different users under several typical lighting conditions: indoor, outdoor (under different types of weather and time duration of a day), fluorescent, incandescent, and shadowed (both uniform and nonuniform cases are considered). Moreover, we capture the images using eight types of popular smartphones which are listed in the following table.
| ID | Model Name | Megapixels (MP) | Image Stabilization | Autofocus |
|---|---|---|---|---|
| 1 | iPhone 6 Plus | 8.0 | ✓ | ✓ |
| 2 | iPhone 6 | 8.0 | ✗ | ✓ |
| 3 | Nexus 5 | 8.0 | ✓ | ✓ |
| 4 | Nexus 4 | 8.0 | ✗ | ✓ |
| 5 | Meizu MX2 | 8.0 | ✗ | ✓ |
| 6 | Oneplus 1 | 13.0 | ✗ | ✓ |
| 7 | Galaxy Nexus 3 | 8.0 | ✗ | ✓ |
| 8 | Sony Xperia M2 | 8.0 | ✗ | ✓ |
News
- 2015-10-14 CUHC-CQRC website is online available. Currently, CUHK_CQRC is still under our final adjustment, and will be coming soon.
References
If you use the CUHK-CQRC database, please cite as:
- Z. Yang, Z. Cheng, C. C. Loy, W. C. Lau, C. M. Li, and G. Li. Towards Robust Color Recovery for High-Capacity Color QR Codes. In Image Processing (ICIP), 2016 IEEE International Conference on, pages 2866–2870. IEEE, 2016.
- Z. Yang, H. Xu, J. Deng, C. C. Loy and W. C. Lau. Robust and Fast Decoding of High-Capacity Color QR Codes for Mobile Applications. In IEEE Transactions on Image Processing, vol. 27, no. 12, pp. 6093-6108, Dec. 2018.
Note
- You agree not to reproduce, duplicate, copy, sell, trade, resell or exploit for any commercial purposes, any portion of the images and any portion of derived data.
- You agree not to further copy, publish or distribute any portion of the CUHK-CQRC dataset. Except, for internal use at a single site within the same organization it is allowed to make copies of the database.
- The MobiTec@CUHK reserves the right to terminate your access to the database at any time.
Download Instruction
Click Here to download.