Face recognition using the nearest feature line method

Abstract

In this paper, we propose a novel classification method, called the nearest feature line (NFL), for face recognition. Any two feature points of the same class (person) are generalized by the feature line (FL) passing through the two points. The derived FL can capture more variations of face images than the original points and thus expands the capacity of the available database. The classification is based on the nearest distance from the query feature point to each FL. With a combined face database, the NFL error rate is about 43.7-65.4% of that of the standard eigenface method. Moreover, the NFL achieves the lowest error rate reported to date for the ORL face database.

Neighbor Line-Based Locally Linear Embedding

“…For example, in the Smote algorithm [1], virtual samples of the minority class are generated such that the number of minority training samples is increased. Here the virtual samples are generated through interpolating between each minority class point and its k nearest neighboring minority class points, which looks somewhat like the interpolating scheme used in Nfl [8] and Nnl [13]. In fact, the Nfl and Nnl algorithms have implicitly utilized virtual samples since they use a virtual point instead of a real data point to help compute the distance between a data point and a class.…”

“…The Nfl (Nearest Feature Line) method was originally proposed for face recognition [8]. In Nfl, feature line is defined as the line passing through two points from a same class.…”

Image recognition via two-dimensional random projection and nearest constrained subspace

“…Classical but still popular subspace-based classifiers include NN (Nearest Neighbor), NFL (Nearest Feature Line, proposed by Stan Z. Li et al [35]) and NS (Nearest Subspace).…”

Robust Kernel Representation With Statistical Local Features for Face Recognition

“…Different from Nearest Neighbor (NN) and Nearest Subspace (NS) classifiers [19] [36][37][38][51] [58], which forbids representing the query sample across classes, the recently developed l 1 -regularized sparse representation [10] or l 2 -regularized collaborative representation [33] represents the query image by the training samples from all classes, which could effectively overcome the small-sample-size or overfitting problem of NN and NS. Let ⋅ is the l p -norm, and p=1 for SRC in [10], while p=2 for CRC in [33].…”

“…Moreover, in order to better exploit the prior knowledge that face images from the same subject construct a subspace, nearest subspace (NS) classifiers [19] [36][37][38][51] [58] were also developed, which are usually superior to the popular NN classifier. Recently an interesting classifier, namely sparse representation based classification (SRC), was proposed by Wright et al [10] for robust FR.…”