Deep Learning Based Tongue Prickles Detection in Traditional Chinese Medicine

Algorithm 1

Mask morphology processing.

	Input: Binary tongue mask I width and height H
	Output: Tongue mask with structural integrity
	num_label = 0
	for y = 0 to H − 1 do
	for x = 0 to − 1 do
	N = neighbours (I[y, x])
	if I[y, x] = = 1 then
	if 0 not in N then
	I[y, x] = num_label//Give each mask pixel a label
	num_label+=1
	else if has_mask(N) then
	label = min(N)
	I[y, x] = label//If neighbours have label, use the minimum one.
	labelSet[label].append(N.all_labels())
	for y = 0 to H − 1 do
	for x = 0 to − 1 do
	I(y, x) = min(labelSet[I(y, x)])//Unify the label of each component
	Select the mask component with largest area as tongue and discard others
	Select the background component with largest area and discard others//Fill holes in mask
	Return I