Efficient Deep Learning Architecture for Detection and Recognition of Thyroid Nodules
Algorithm 1
K-means get anchor box.
Input: gtbox, K.
gtbox is a set of ground truth boxes. gtbox = {gtbox(1), ..., gtbox(n)}. gtbox(i) = (i, hi), ∀I ∈ [1, n]
Output: C, abox.
C = {C1, ..., CK} represents the cluster center coordinates of K categories, respectively. abox is a set of K boxes of anchor boxes. abox = {abox1, ..., gtboxk}. aboxk = (k, hk), ∀k ∈ [1, K]
(1)
for k = 0 − > K do
(2)
C < −Random(CK)
(3)
end for
(4)
/NewC = {NewC1, ..., NewCK} represents the updated cluster center coordinates of K categories, respectively./
(5)
NewC < −NULL
(6)
while NewC ! = C do
(7)
for i = 0 − > n do
(8)
μ(i) = arg min d(getbox(i)-CK)/μ = μ(1), ..., μ(n) is the index of the cluster center closest to gtbox(i)∗/
