Algorithm 1: Landmark Detection and Feature Extraction

Input: Upper facial image Output: Landmark coordinate and HOG feature landmarks t $\leftarrow$ $\left[0,\frac{20\mathsf{\pi }}{50},\frac{2\times 20\mathsf{\pi }}{50},\frac{3\times 20\mathsf{\pi }}{50},\dots ,20\mathsf{\pi }\right]$ x $\leftarrow$ $\cos \mathrm{t}$ y $\leftarrow$ $\frac{\sin 4\mathrm{t}}{2}$ points $\leftarrow$ (x, y) for i=0 to (length of t) − 3 do    startPoint $\leftarrow$ points[i]    midPoint $\leftarrow$ points[i+1]    endPoint $\leftarrow$ points[i+2]    connectedLineIntensity[i] $\leftarrow$ connect(startPoint, midPoint, endPoint) end for height $\leftarrow$ height of peak fwhm $\leftarrow$ full width of half maximum for each lineIntensity in connectedLineIntensity do    (indices, properties) $\leftarrow$ findPeaks(lineIntensity) end for for i=0 to length of properties do    if prominence of properties[i] >= height     and width of properties[i]>=fwhm then      property $\leftarrow$ properties[i]      landmarks[i] $\leftarrow$ property[leftBase]    end end for for i=0 to length of landmarks do   landmarkHOG[i] $\leftarrow$ histogramofOrientedGradient(landmarks[i])   features[i] $\leftarrow$ (landmark[i], landmarkHOG[i]) end for