From Local Geometry to Global Structure: Learning Latent Subspace for Low-resolution Face Image Recognition

Abstract

In this letter, we propose a novel approach for learning coupled mappings to improve the performance of low-resolution (LR) face image recognition. The coupled mappings aim to project the LR probe images and high-resolution (HR) gallery images into a unified latent subspace, which is efficient to measure the similarity of face images with different resolutions. In the training phase, we first construct local optimization for each training sample according to the relationship of neighboring data points. The local optimization aims to: (1) ensure the consistency for each LR face image and corresponding HR one; (2) model the intrinsic geometric structure between each given sample and its neighbors; and (3) preserve the discriminative information across different subjects. We finally incorporate the local optimizations together for building the global structure. The coupled mappings can be learned by solving a standard eigen-decomposition problem, which avoids the small-sample-size problem. Experimental results demonstrate the effectiveness of the proposed method on public face databases.

Preclinical Diagnosis of Magnetic Resonance (MR) Brain Images via Discrete Wavelet Packet Transform with Tsallis Entropy and Generalized Eigenvalue Proximal Support Vector Machine (GEPSVM)

“…Co-registration was not needed because it is not essential. The proposed technique was similar to those used for face recognition, in which some scholars used co-registration [18,19] but others did not use it [20,21]. Moreover, some past publications about abnormal brain detection did not use co-registration [5][6][7][8][9][10][11][12][13][14][15] but they all obtained good results, which were comparable to those obtained employing co-registration [22,23].…”

A Study on the Performance of Unconstrained Very Low Resolution Face Recognition: Analyzing Current Trends and New Research Directions

“…The Local Geometry to Global Structure CM approach proposed by Shi et al [86] aims to minimize the distance between projected features as well. However, this approach uses a k-neighbor approach to influence the distance optimization in both intra-class and inter-class projected groups.…”

On Low-Resolution Face Recognition in the Wild: Comparisons and New Techniques

“…We follow the experiment setting provided in [73] and [59] and study both closed-set and open-set scenarios. For closed-set evaluation, 180 subjects are used and for open set evaluation, we compare our result with the performance reported in [59] with the HD-1m HD-2.6m HD-4.7m SCface [14] 6.18% 6.18% 1.82% CLPM [35] 3.08% 4.32% 3.46% SSR [82] 18.09% 13.2% 7.04% CSCDN [75] 18.97% 13.58% 6.99% CCA [76] 20.69% 14.85% 9.79% DCA [16] 25.53% 18.44% 12.19% C-RSDA [11] 18.46% 18.08% 15.77% Centerloss [77] [35] 29.12% SDA [91] 40.08% CMFA [64] 39.56% Coupled mapping method [63] 43.24% LMCM [87] 60.40% Centerloss [77] D. Low-resolution face re-identification In this section, we explore LR face re-identification and evaluate it on several datasets captured in an unconstrained environment. We employ the VBOLO dataset for an in-depth study and the SCface, UCCSface, and MegaFace challenge 2 LR subset for other topical explorations.…”