Video Based Face Recognition by Using Discriminatively Learned Convex Models
- PDF / 1,247,087 Bytes
- 15 Pages / 595.276 x 790.866 pts Page_size
- 56 Downloads / 146 Views
Video Based Face Recognition by Using Discriminatively Learned Convex Models Hakan Cevikalp1
· Golara Ghorban Dordinejad1
Received: 19 July 2019 / Accepted: 8 July 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract A majority of the image set based face recognition methods use a generatively learned model for each person that is learned independently by ignoring the other persons in the gallery set. In contrast to these methods, this paper introduces a novel method that searches for discriminative convex models that best fit to an individual’s face images but at the same time are as far as possible from the images of other persons in the gallery. We learn discriminative convex models for both affine and convex hulls of image sets. During testing, distances from the query set images to these models are computed efficiently by using simple matrix multiplications, and the query set is assigned to the person in the gallery whose image set is closest to the query images. The proposed method significantly outperforms other methods using generatively learned convex models in terms of both accuracy and testing time, and achieves the state-of-the-art results on six of the eight tested datasets. Especially, the accuracy improvement is significant on the challenging PaSC, COX, IJB-C and ESOGU video datasets. Keywords Discriminative models · Affine hulls · Convex hulls · Face recognition · Image sets
1 Introduction Face recognition is an important computer vision problem that has many applications in various fields. Initially, single images are used for face recognition, but more recently, set based methods have begun to dominate the field mostly because face image sets allow to model the variability of the individuals’ appearances. For set based face recognition, both gallery and query sets are given in terms of sets of images rather than a single image. Images can be collected from video frames as well as from multiple unordered observations. The classification system must return the individual whose gallery set is the most similar to the given query set. Face recognition methods using image sets are also more practical owing to the fact that they usually do not require any cooperation from the subjects. However, despite these Communicated by Ming-Hsuan Yang.
B
Hakan Cevikalp [email protected] Golara Ghorban Dordinejad [email protected]
1
Electrical and Electronics Engineering, Machine Learning and Computer Vision Laboratory, Eskisehir Osmagazi University, Eskisehir, Turkey
advantages, traditional classifiers such as the support vector machines (SVMs), classification trees, k-nearest neighbor classifier, etc. cannot be used directly, which can be considered as a major limitation of the set based methods. There are two important factors that determine the success of the set based face recognition methods: the models used to approximate the face image sets, and the distance metric used to measure the similarity between these models. A variety of different models were use
Data Loading...
