Face Classification Based on Linguistic Description of Facial Features

A study in facial features saliency in face recognition: an analytic hierarchy process approach

“…In many experimental studies concerning the way of the process of recognition by people or by computers, the authors assume the obligatory partition of the face, e.g., upper and lower half of the face (Haig 1986), forehead area (including the eyes), nose region (including cheeks and ears), and the mouth region (including chin, cf. Kurach et al 2014), the areas of eyebrows, eyes, nose, mouth, cheeks (Karczmarek et al 2014), the regions of eyebrows, eyes, nose, mouth, chin, and hair (Matthews 1978), and other partitions.…”

“…Furthermore, global precedence hypothesis corresponds to the paradigm of granular computing (Pedrycz 2013). Therefore, the facial features can be grouped into meaningful and semantically sound entities, referred to information granules such as internal/external facial features (e.g., eyes, and nose/chin and ears), upper/lower half of a face, and eyes/nose/mouth areas (the last partition was described by Kurach et al 2014). Of course, each of these general information granules consists of "atomic" facial parts.…”

Linguistic Descriptors in Face Recognition

“…Finally, the linguistic descriptors were measured by using experts' votes in [21]. Other approaches based on linguistic descriptors expressed in terms of fuzzy sets, fuzzy geometries, granular computing, and others were described in [22][23][24][25][26][27][28]. A comprehensive survey of methods utilizing the linguistic descriptors in face recognition can be found in [29].…”

An application of chain code-based local descriptor and its extension to face recognition

“…There are many trends in face recognition. The most significant and comprehensively examined are the approaches originated from geometrical methods [1], principal component analysis (eigenfaces [2]), linear discriminant analysis (Fisherfaces [3]), EBGM (elastic bunch graph matching [4]), SVMs (support vector machines [5]), Gabor wavelets [6,7], information aggregation and fusion [8], neural networks [9], sparse representation [10], deep learning [11,12], Granular Computing and linguistic descriptors [13,14], and local descriptors [15,16].For instance, the latter class of methods is,in general, invariant to lighting and pose changes. Other advantages of local descriptors are their relatively low computational cost and very intuitive processing scheme taking as an input the neighborhood of a given pixel and resulting in the value being the description of this surrounding.…”

“…between rank 1 recognition rates and the dictionary length for the CCBLD with block size p=7 84 13. 87.08 90.10 91.50 91.85 92.78 92.55 92.78 93.28 94.03 94.68 95.18 95.28 2x2 92.18 94.65 95.88 96.45 96.58 96.78 96.85 97.15 97.20 97.05 97.48 97.65 97.53 97.70 3x3 96.33 97.35 97.73 97.93 98.00 98.15 98.13 98.10 .28 4x4 96.65 97.70 97.80 97.73 97.83 97.85 97.68 98.05 97.88 98.05 97.88 97.90 97.88 97.95 5x5 96.18 96.85 96.80 97.03 96.95 97.23 97.03 97.23 97.18 97.20 97.18 97.15 97.05 96.95 6x6 96.60 96.68 96.75 96.88 96.60 96.65 96.65 96.70 96.68 96.58 96.58 96.33 96.38 96.50 7x7 95.80 96.13 96.10 96.38 96.30 96.35 96.25 96.55 96.28 96.35 96.33 96.28 96.43 96.40 8x8 95.65 95.88 95.85 96.13 96.13 96.15 96.28 96.30 95.98 96.05 96.18 96.18 96.23 96.33 9x9 95.60 95.83 95.70 95.70 95.90 95.78 95.85 95.48 95.65 95.65 95.73 95.60 95.70 95.63 10x10 95.08 95.55 95.75 95.63 95.58 95.50 95.58 95.40 95.45 95.58 95.35 95.55 95.40 95.40…”