Using the original and 'symmetrical face' training samples to perform representation based two-step face recognition
Citations
171 citations
Cites background or methods from "Using the original and 'symmetrical..."
...However, as shown in [18], [20], and [24], by exploiting a special means, l2 regularization can also obtain satisfactory sparse property....
[...]
...For example, both robust regression for classification [23] and two-phase test sample sparse representation [20], [24] are l2 regularization-based representation methods....
[...]
...Especially, the elaborated algorithms of l2 regularization-based representation, such as the ones designed in [20] and [24], can obtain quite sparse representation coefficients, and flexibly control the extent of sparsity....
[...]
111 citations
Cites background or methods from "Using the original and 'symmetrical..."
...8, one can see that our method in most cases can achieve higher classification accuracies than LCKSVD, DKSVD, SFTS, MFTS, ICR, MFL and RNVS....
[...]
...For example, symmetrical face images proposed in [11] are very beneficial to overcome the problem of varying poses and illuminations....
[...]
...For example, symmetrical face training samples (SFTS) [11], inter-class relationship (ICR) [13], random noise based virtual sample (RNVS) [14] and mirror face training samples (MFTS) [12] were all proposed for face recognition....
[...]
...SFTS Symmetrical face training samples [11] MFTS Mirror face training samples [3] ICR Inter-class relationship [13] MFL Metaface learning [59] RNVS Random noise based virtual sample [14]...
[...]
86 citations
Cites background or methods or result from "Using the original and 'symmetrical..."
...8 and 9 show some reasonable and unreasonable “symmetrical” face images obtained in [15]....
[...]
...However, it should be pointed out that in real-world face recognition applications, a large number of face images are not symmetrical images due to non-frontal and non-neutral pose [15]....
[...]
...As shown in [15], these virtual face images can somewhat reflect possible variation in pose and scale of the face....
[...]
...The scheme proposed in [15] is able to obtain a virtual axis-symmetrical face image....
[...]
...proposed an approach to generate “symmetrical” face images and exploited both the original and “symmetrical” face images to recognize the subject [15]....
[...]
78 citations
60 citations
Cites background from "Using the original and 'symmetrical..."
...Multiple virtual views are synthesized by linear shape prediction [29], mesh warping [30], morphing [31], symmetry property [32], partitioning a face in several sub-images [33], affine transformation [34], noise perturbation [35], shifting [36], and active appearance model [37]....
[...]
References
12,972 citations
"Using the original and 'symmetrical..." refers methods in this paper
...4 shows a test sample (4(a)) that is erroneously and correctly classified by the collaborative representation (CR) method proposed in [47] and our method, respectively....
[...]
...Besides our proposed method was tested, several state-of-the-art face recognition methods such as two-phase test sample sparse representation (TPTSSR) [45], collaborative representation (CR) proposed in [47], the feature space-based representation method proposed in [49], sparse representation classification (SRC) proposed in [50] and linear regression classification (LRC) [51]were also tested....
[...]
9,658 citations
3,439 citations
2,954 citations
"Using the original and 'symmetrical..." refers background in this paper
...However, up to now, face recognition is still faced with a number of challenges such as varying illumination, facial expression and poses [10–16]....
[...]
2,001 citations