A Multi-Face Challenging Dataset for Robust Face Recognition

  • 2018-09-30 07:04:59
  • Shiv Ram Dubey, Snehasis Mukherjee
  • 19


Face recognition in images is an active area of interest among the computervision researchers. However, recognizing human face in an unconstrainedenvironment, is a relatively less-explored area of research. Multiple facerecognition in unconstrained environment is a challenging task, due to thevariation of view-point, scale, pose, illumination and expression of the faceimages. Partial occlusion of faces makes the recognition task even morechallenging. The contribution of this paper is two-folds: introducing achallenging multiface dataset (i.e., IIITS MFace Dataset) for face recognitionin unconstrained environment and evaluating the performance of state-of-the-arthand-designed and deep learning based face descriptors on the dataset. Theproposed IIITS MFace dataset contains faces with challenges like posevariation, occlusion, mask, spectacle, expressions, change of illumination,etc. We experiment with several state-of-the-art face descriptors, includingrecent deep learning based face descriptors like VGGFace, and compare with theexisting benchmark face datasets. Results of the experiments clearly show thatthe difficulty level of the proposed dataset is much higher compared to thebenchmark datasets.


Introduction (beta)



Conclusion (beta)