<title>Semantic segmentation of videophone image sequences</title>

Face Detection: A Survey

“…For instance, a small area on top of a larger area in a head and shoulder sequence implies a "face on top of shoulder" scenario, and a pair of dark regions found in the face area increase the confidence of a face existence. Among the literature survey, a pair of eyes is the most commonly applied reference feature [5,27,52,61,207,214] due to its distinct side-by-side appearance. Other features include a main face axis [26,165], outline (top of the head) [26,32,162] and body (below the head) [192,207].…”

“…This property can be exploited to differentiate various facial parts. Several recent facial feature extraction algorithms [5,53,100] search for local gray minima within segmented facial regions. In these algorithms, the input images are first enhanced by contrast-stretching and gray-scale morphological routines to improve the quality of local dark patches and thereby make detection easier.…”

“…This approach, whilst simple, is able to discern a moving foreground efficiently regardless of the background content. In [5,53,152,192,218], moving silhouettes that include face and body parts are extracted by thresholding accumulated frame difference. Besides face region, Luthon and Lievin [118], Crowley and Berard [27], and Low [115,116] also employ frame difference to locate facial features.…”

Human Face Detection Techniques: A Comprehensive Review and Future Research Directions

“…Feature-based approaches for face detection: This can be broadly classified into active shape model [10][11][12][13][14][15][16][17][18][19][20][21], low level analysis [22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37], and feature analysis [38][39][40][41][42][43][44][45][46][47][48][49][50][51][52][53][54][55][56]. Each of these again can be classified into several subcategories as shown here.…”

“…In frame variance analysis, the moving forepart is identified in any type of background. Moving parts that contain a face are discerned by thresholding the gathered frame difference [22,23]. Along with the face region, face features can also be extracted in this way [24,25].…”