An eyeglass face dataset collected and cleaned for face recognition evaluation.
Branch: master
Clone or download
Latest commit 1c91dbb Dec 4, 2018
Type Name Latest commit message Commit time
Failed to load latest commit information.
samples Update and samples Jun 6, 2018
test Update download link and Jun 5, 2018
LICENSE Initial commit Jun 2, 2018 Update Dec 4, 2018
meta.txt Update download link and Jun 5, 2018

Face Synthesis for Eyeglass-Robust Face Recognition



This repo releases the MeGlass dataset in original paper. MeGlass is an eyeglass dataset originaly designed for eyeglass face recognition evaluation. All the face images are selected and cleaned from MegaFace. Each identity has at least two face images with eyeglass and two face images without eyeglass. More details are presented in paper Face Synthesis for Eyeglass-Robust Face Recognition.

Name Dataset type Link Cropped Google Drive or Baidu Yun, 335.8M Origin Baidu Yun, 13.3G

Dataset description

meta.txt contains the eyeglass labels of images. 1 means black-eyeglass, 0 means no-eyeglass. consists of the cropped images of size 120x120. contains the original face images.

test directory contains four lists corresponding to the four protocols in paper.

Dataset Identity Images Black-eyeglass No-eyeglass
MeGlass 1,710 47,917 14,832 33,085
Testing set 1,710 6,840 3,420 3,420


Dataset usages

To build this dataset, we use eyeglass classifier, powerful face recognition model and manual labor to keep right the person identity and black eyeglass attribute. Therefore, MeGlass dataset can be used for face recognition (identification and verification), eyeglass detection, removal, generation tasks and so on.

Identity parsing rule

Take one filename 10032527@N08_identity_4@2897031059_1.jpg for example, the string before the second @ makes one face image's identity. The naming rule is corresponding to the original MegaFace dataset.


The 3D face model fitting is based on Xiangyu Zhu's work.


If your research benefits from MeGlass, please cite it as

  title={Face Synthesis for Eyeglass-Robust Face Recognition},
  author={Guo, Jianzhu and Zhu, Xiangyu and Lei, Zhen and Li, Stan Z},
  journal={arXiv preprint arXiv:1806.01196},