Human Pathogenic Monkeypox Disease Recognition Using Q-Learning Approach

. 2023 Apr 20;13(8):1491. doi: 10.3390/diagnostics13081491

Algorithm 1: Malsneural algorithm

1. Procedure Augmentation(image, pro)
2. prob pro:
3.

i m a g e \leftarrow R o t a t e (i m a g e, (- 5, + 5))

4. prob pro:
5.

i m a g e \leftarrow T r a n s l a t e (i m a g e, (0.8, 1.2))

6. prob pro:
7.

i m a g e \leftarrow S a t u r a t i o n (i m a g e)

8. prob pro:
9.

i m a g e \leftarrow S c a l i n g (i m a g e, (0.7, 1.2))

10. prob pro:
11.

i m a g e \leftarrow h u e (i m a g e)

12. Return image
13. Adaptive median filter
14. Level 1:
15.

i m a g e 1 = Z_{m e d i a n} - Z_{m i n}

16.

i m a g e 2 = Z_{m e d i a n} - Z_{m a x}

17. If image1 > 0 and image 2 < 0 go to the next level
18. Else the size of the window increased
19. If windoe size <= size max redo the level 1
20. Else return zxy
21. Level 2:
22.

i m a g e 3 = Z_{x y} - Z_{m i n}

23.

i m a g e 4 = Z_{x y} - Z_{m a x}

24. If image 3 > 0 and image 4 < 0 return zxy
25. Else return zmedian
26. End if
27. Load replay memory M to the capacity C
28. Load the function action Q along with arbitrary weight W
29. Load destination value function Q along with weight W- = W
30. For iteration = 1,N do
31. Load sequence t = {y1} and preprocessed ϕ1 = ϕ(t1)
32. For q = 1, Q do
33. The random action choosen bQ
34. Orelse choose bq = argmaxb P(ϕ(tq),b;W)
35. Compile bq in emulator and notice reward rq and yq + 1 of input
36. Set t q + 1 = tq, bq, y q + 1 and process ϕq + 1 = ϕ(tq + 1)
37. Save the transition (ϕq, bq,rq,ϕq + 1) in M
38. Minibatch (ϕi, bi, fi,ϕi + 1 ) from M
39. If it stops at i + 1
40. Initialise fj
41. Else
42. Yj = {fi + ϑmax d P(ϕi + 1,bq,W)
43. Execute gradient descent by updating the gradient value (yi-P(ϕi,bi; W))2
44. Reset ό = P
45. End for
46. End for