Algorithm 1: Adaptive Stochastic Gradient Descent Algorithm

Input: Fetal abnormality images.

Output: Risk prediction

start:

# Remove the noise in the fetal abnormality images

Do

training data classification

for j = 1 to Data

initnumpy (array) → size

[A,B,C]

returns: [(class image)]

returns: [corres, m(images)]

end for

pat_size = array(array, shape)

class = trans (gauss_noise)

m = trans (spat)

m = m [0]

return (class [0], m)

for each i in data

returns: sample(w)

n_samples = l(s_labels)

CountDiC = dict(unique, Count)

w = [ ]

for each label in s_labels

Append (w) → (n_samples/count labels)

return w

end for

end for

update w

while (iter<= imagei)

End