github/DeepFaceLab
1
0
Fork 0
mirror of https://github.com/iperov/DeepFaceLab.git synced 2025-07-06 21:12:07 -07:00
Code Issues Projects Releases Packages Wiki Activity Actions

new fix of "fixed bug when the same face could be detected twice"

Browse source
This commit is contained in:
Colombo 2019-10-14 20:26:53 +04:00
parent 85cbdcd290
commit e18b07549b
1 changed files with 29 additions and 15 deletions
Download patch file Download diff file Expand all files Collapse all files

36
facelib/FANExtractor.py
View file

@ -4,6 +4,7 @@ from pathlib import Path
import cv2 import cv2
import numpy as np import numpy as np
from numpy import linalg as npla
from facelib import FaceType, LandmarksProcessor from facelib import FaceType, LandmarksProcessor
from nnlib import nnlib from nnlib import nnlib
@ -76,24 +77,37 @@ class FANExtractor(object):
landmarks.append (None) landmarks.append (None)
if second_pass_extractor is not None: if second_pass_extractor is not None:
for i in range(len(landmarks)): new_landmarks = []
try:
lmrks = landmarks[i]
if lmrks is None:
continue
for lmrks in landmarks:
new_landmarks += [None]
try:
if lmrks is not None:
image_to_face_mat = LandmarksProcessor.get_transform_mat (lmrks, 256, FaceType.FULL) image_to_face_mat = LandmarksProcessor.get_transform_mat (lmrks, 256, FaceType.FULL)
face_image = cv2.warpAffine(input_image, image_to_face_mat, (256, 256), cv2.INTER_CUBIC ) face_image = cv2.warpAffine(input_image, image_to_face_mat, (256, 256), cv2.INTER_CUBIC )
rects2 = second_pass_extractor.extract(face_image, is_bgr=is_bgr) rects2 = second_pass_extractor.extract(face_image, is_bgr=is_bgr)
if len(rects2) != 1: #dont do second pass if faces != 1 detected in cropped image if len(rects2) == 1: #dont do second pass if faces != 1 detected in cropped image
continue
lmrks2 = self.extract (face_image, [ rects2[0] ], is_bgr=is_bgr, multi_sample=True)[0] lmrks2 = self.extract (face_image, [ rects2[0] ], is_bgr=is_bgr, multi_sample=True)[0]
source_lmrks2 = LandmarksProcessor.transform_points (lmrks2, image_to_face_mat, True) new_landmarks[-1] = LandmarksProcessor.transform_points (lmrks2, image_to_face_mat, True)
landmarks[i] = source_lmrks2
except: except:
continue pass
for i, new_lmrks in enumerate(new_landmarks):
mat = LandmarksProcessor.get_transform_mat (new_lmrks, 256, FaceType.FULL)
center, p1,p2 = LandmarksProcessor.transform_points ([ [127,127], [0,0], [0,255] ], mat, True)
p_dist = npla.norm(p2-p1)
for j, other_lmrks in enumerate(landmarks):
if i != j:
other_mat = LandmarksProcessor.get_transform_mat (new_lmrks, 256, FaceType.FULL)
other_center = LandmarksProcessor.transform_points ([ [127,127] ], other_mat, True)
dist = npla.norm (other_center - center)
if dist < p_dist*0.25:
break
else:
landmarks[i] = new_lmrks
return landmarks return landmarks