Entropy based Binary Particle Swarm Optimization and classification for ear detection

Entropy-cum-Hough-transform-based ear detection using ellipsoid particle swarm optimization

“…Owing to the use of different datasets, a quantitative comparison of performance could not be made with [34] and [24]. Madan et al [28] have used an entropy-based classification. Unlike our method, they use binary particle swarm optimization to optimize an entropy map, which is then used to localize the ear.…”

“…Their technique involves considerable amount of time for training and the detection itself takes an average time of 12.18 s. Entropy is just a mathematical measure and in real-time scenarios, cases with other extensive feature-rich areas around the ear are often encountered. Any feature with an equivalent or higher entropy than the ear might be falsely accepted as an ear by the classifier in [28]. Their method also needs prior knowledge of the face profile for detection and is not fully automated and thus unsuitable for real-time scenarios.…”

“…Ganesh et al [28] have used an entropy-based classification, where an entropy map is created and binary particle swarm optimization is used to optimize this map, which is then used to localize the ear. Their detection is primarily based on the observed patterns of entropy.…”

Ear Detection under Uncontrolled Conditions with Multiple Scale Faster Region-Based Convolutional Neural Networks

“…However, these ear extraction methods only work on profile images without any kind of rotation, pose, or scale variation and occultation. Ganesh and Krishna proposed an approach to detect ears in facial images under uncontrolled environments [24]. They proposed at echnique, namely Entropic Binary Particle Swarm Optimization (EBPSO), which generated an entropy map, the highest value of which was used to localize the ear in a face image.…”

“…Detection Approach Dataset Size Accuracy Rate % Burge and Burger [10] deformable contours N/A N/A Hurley et al [11] force filed transform N/A N/A Alvarez et al [12] the snake model and ovoid model N/A N/A Ansari and Gupta [13] Edge Detection and Curvature Estimation 700 93.34 Yuan and Mu [14] skin-color model and intensity contour information Video N/A Arbab [15] Hough Transform 942 91 Cummings et al [16] the image ray transform 252 98.4 Prakash and Gupta [17] skin color and Graph matching 1780 96.63 Yan and Bowyer [18] two-line landmark N/A N/A Yan and Bowyer [19] ear pit detection and active contour 415 78.8/85.54 Deepak et al [20] Active contour model 700 76.43 Chen and Bbanu [21] the step edge magnitude N/A N/A Chen and Bbanu [22] the step edge and ear shape model 312 92.6 Chen and Bbanu [23] skin classifier, the step edge and ear shape model 700 87.71 Ganesh et al [24] EBPSO and DTCWT 240 92.92 Sana et al [25] ear templates of different sizes N/A N/A Prakash et al [26] skin-color classifier and template matching 150 94 Attarchi et al [27] the edge map and the mean ear template 308 98.05 Halawani [28] predefined binary ear template 212 96.2 Joshi [29] oval shape detection 798 94% Islam et al [30] cascaded AdaBoost 203 100 Abaza et al [31] modified AdaBoost >2000 95 Yuan and Mu [33] improved AdaBoost 434 95.8…”