Fix【wrong type】about eyesCenter

[BeGentleman]

Error Info:【TypeError: Can't parse 'center'. Sequence item with index 0 has a wrong type】
Cause by: eyesCenter's element's type is numpy.int64, which is need an int type
Solved by: change numpy.int64 into int type
sources: eyesCenter = ((leftEyeCenter[0] + rightEyeCenter[0]) // 2, (leftEyeCenter[1] + rightEyeCenter[1]) // 2) ->
eyesCenter = (int((leftEyeCenter[0] + rightEyeCenter[0]) // 2), int((leftEyeCenter[1] + rightEyeCenter[1]) // 2))

[AiDeepAlgo]

I want to verify this fix, I have tested this.
I'm also asking if this can be marge and integrated into a new release.
thanks

[mlgarrett]

i too would love to see this fix merged as soon as possible! i have tested it as well and casting to int does make it work.

[mcp292]

Fixes #278.

[iaglourenco]

Any possibility to merge this PR?

[Darkiee12]

Does this fix work? I believe we need the cast to float instead here

classFaceAligner:
defalign(self, image, gray, rect):
eyesCenter: float= ((leftEyeCenter[0] +rightEyeCenter[0]) /2, # change from // to / to get float value
(leftEyeCenter[1] +rightEyeCenter[1]) /2) # do the sameM=cv2.getRotationMatrix2D(eyesCenter, angle, scale)

As this function calls the C++ written function in OpenCV, stated here,

cv::getRotationMatrix2D(Point2f center, double angle, double scale)
typedef Point_<float> cv::Point2f