github/DeepFaceLab
1
0
Fork 0
mirror of https://github.com/iperov/DeepFaceLab.git synced 2025-08-22 06:23:20 -07:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge pull request #27 from faceshiftlabs/feat/consistent_dpi

Browse source 
Feat/consistent dpi
This commit is contained in:
Jeremy Hummel 2019-08-23 14:04:17 -07:00 • committed by GitHub
commit 2add4e2bad
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
6 changed files with 58 additions and 45 deletions
Download patch file Download diff file Expand all files Collapse all files

1
facelib/LandmarksProcessor.py
View file

@ -149,6 +149,7 @@ def get_transform_mat (image_landmarks, output_size, face_type, scale=1.0):
else:
raise ValueError ('wrong face_type: ', face_type)
mat = umeyama(image_landmarks[17:], landmarks_2D, True)[0:2]
mat = mat * (output_size - 2 * padding)
mat[:,2] += padding

61
facelib/S3FDExtractor.py
View file

@ -3,26 +3,39 @@ from pathlib import Path
import cv2
from nnlib import nnlib
class S3FDExtractor(object):
"""
S3FD: Single Shot Scale-invariant Face Detector
https://arxiv.org/pdf/1708.05237.pdf
"""
def __init__(self):
exec( nnlib.import_all(), locals(), globals() )
exec(nnlib.import_all(), locals(), globals())
model_path = Path(__file__).parent / "S3FD.h5"
if not model_path.exists():
return None
raise Exception(f'Could not find S3DF model at path {model_path}')
self.model = nnlib.keras.models.load_model ( str(model_path) )
self.model = nnlib.keras.models.load_model(str(model_path))
def __enter__(self):
return self
def __exit__(self, exc_type=None, exc_value=None, traceback=None):
return False #pass exception between __enter__ and __exit__ to outter level
return False # pass exception between __enter__ and __exit__ to outter level
def extract (self, input_image, is_bgr=True):
def extract(self, input_image, is_bgr=True, nms_thresh=0.3):
"""
Extracts the bounding boxes for all faces found in image
:param input_image: The image to look for faces in
:param is_bgr: Is this image in OpenCV's BGR color mode, if not, assume RGB color mode
:param nms_thresh: The NMS (non-maximum suppression) threshold. Of all bounding boxes found, only return
bounding boxes with an overlap ratio less then threshold
:return:
"""
if is_bgr:
input_image = input_image[:,:,::-1]
input_image = input_image[:, :, ::-1]
is_bgr = False
(h, w, ch) = input_image.shape
@ -32,35 +45,36 @@ class S3FDExtractor(object):
scale_to = max(64, scale_to)
input_scale = d / scale_to
input_image = cv2.resize (input_image, ( int(w/input_scale), int(h/input_scale) ), interpolation=cv2.INTER_LINEAR)
input_image = cv2.resize(input_image, (int(w / input_scale), int(h / input_scale)),
interpolation=cv2.INTER_LINEAR)
olist = self.model.predict( np.expand_dims(input_image,0) )
olist = self.model.predict(np.expand_dims(input_image, 0))
detected_faces = []
for ltrb in self.refine (olist):
l,t,r,b = [ x*input_scale for x in ltrb]
bt = b-t
if min(r-l,bt) < 40: #filtering faces < 40pix by any side
for ltrb in self._refine(olist, nms_thresh):
l, t, r, b = [x * input_scale for x in ltrb]
bt = b - t
if min(r - l, bt) < 40: # filtering faces < 40pix by any side
continue
b += bt*0.1 #enlarging bottom line a bit for 2DFAN-4, because default is not enough covering a chin
detected_faces.append ( [int(x) for x in (l,t,r,b) ] )
b += bt * 0.1 # enlarging bottom line a bit for 2DFAN-4, because default is not enough covering a chin
detected_faces.append([int(x) for x in (l, t, r, b)])
return detected_faces
def refine(self, olist):
def _refine(self, olist, thresh):
bboxlist = []
for i, ((ocls,), (oreg,)) in enumerate ( zip ( olist[::2], olist[1::2] ) ):
stride = 2**(i + 2) # 4,8,16,32,64,128
for i, ((ocls,), (oreg,)) in enumerate(zip(olist[::2], olist[1::2])):
stride = 2 ** (i + 2) # 4,8,16,32,64,128
s_d2 = stride / 2
s_m4 = stride * 4
for hindex, windex in zip(*np.where(ocls > 0.05)):
score = ocls[hindex, windex]
loc = oreg[hindex, windex, :]
loc = oreg[hindex, windex, :]
priors = np.array([windex * stride + s_d2, hindex * stride + s_d2, s_m4, s_m4])
priors_2p = priors[2:]
box = np.concatenate((priors[:2] + loc[:2] * 0.1 * priors_2p,
priors_2p * np.exp(loc[2:] * 0.2)) )
priors_2p * np.exp(loc[2:] * 0.2)))
box[:2] -= box[2:] / 2
box[2:] += box[:2]
@ -69,12 +83,11 @@ class S3FDExtractor(object):
bboxlist = np.array(bboxlist)
if len(bboxlist) == 0:
bboxlist = np.zeros((1, 5))
bboxlist = bboxlist[self.refine_nms(bboxlist, 0.3), :]
bboxlist = [ x[:-1].astype(np.int) for x in bboxlist if x[-1] >= 0.5]
bboxlist = bboxlist[self._refine_nms(bboxlist, thresh), :]
bboxlist = [x[:-1].astype(np.int) for x in bboxlist if x[-1] >= 0.5]
return bboxlist
def refine_nms(self, dets, thresh):
def _refine_nms(self, dets, nms_thresh):
keep = list()
if len(dets) == 0:
return keep
@ -93,6 +106,6 @@ class S3FDExtractor(object):
width, height = np.maximum(0.0, xx_2 - xx_1 + 1), np.maximum(0.0, yy_2 - yy_1 + 1)
ovr = width * height / (areas[i] + areas[order[1:]] - width * height)
inds = np.where(ovr <= thresh)[0]
inds = np.where(ovr <= nms_thresh)[0]
order = order[inds + 1]
return keep

0
guilib/main.py Normal file
View file

4
imagelib/warp.py
View file

@ -4,8 +4,8 @@ from utils import random_utils
def gen_warp_params (source, flip, rotation_range=[-10,10], scale_range=[-0.5, 0.5], tx_range=[-0.05, 0.05], ty_range=[-0.05, 0.05] ):
h,w,c = source.shape
if (h != w) or (w != 64 and w != 128 and w != 256 and w != 512 and w != 1024):
raise ValueError ('TrainingDataGenerator accepts only square power of 2 images.')
#if (h != w) or (w != 64 and w != 128 and w != 256 and w != 512 and w != 1024):
# raise ValueError ('TrainingDataGenerator accepts only square power of 2 images.')
rotation = np.random.uniform( rotation_range[0], rotation_range[1] )
scale = np.random.uniform(1 +scale_range[0], 1 +scale_range[1])

28
mainscripts/Extractor.py
View file

@ -41,7 +41,6 @@ class ExtractSubprocessor(Subprocessor):
#override
def on_initialize(self, client_dict):
self.type = client_dict['type']
self.image_size = client_dict['image_size']
self.face_type = client_dict['face_type']
self.device_idx = client_dict['device_idx']
self.cpu_only = client_dict['device_type'] == 'CPU'
@ -126,6 +125,8 @@ class ExtractSubprocessor(Subprocessor):
else:
h, w, ch = image.shape
#self.image_size = h
if ch == 1:
image = np.repeat (image, 3, -1)
elif ch == 4:
@ -230,6 +231,8 @@ class ExtractSubprocessor(Subprocessor):
continue
rect = np.array(rect)
rect_area = mathlib.polygon_area(np.array(rect[[0, 2, 2, 0]]), np.array(rect[[1, 1, 3, 3]]))
self.image_size = int(math.sqrt(rect_area))
if self.face_type == FaceType.MARK_ONLY:
image_to_face_mat = None
@ -243,7 +246,7 @@ class ExtractSubprocessor(Subprocessor):
landmarks_bbox = LandmarksProcessor.transform_points ( [ (0,0), (0,self.image_size-1), (self.image_size-1, self.image_size-1), (self.image_size-1,0) ], image_to_face_mat, True)
rect_area = mathlib.polygon_area(np.array(rect[[0,2,2,0]]), np.array(rect[[1,1,3,3]]))
landmarks_area = mathlib.polygon_area(landmarks_bbox[:,0], landmarks_bbox[:,1] )
if landmarks_area > 4*rect_area: #get rid of faces which umeyama-landmark-area > 4*detector-rect-area
@ -305,10 +308,9 @@ class ExtractSubprocessor(Subprocessor):
return data.filename
#override
def __init__(self, input_data, type, image_size=None, face_type=None, debug_dir=None, multi_gpu=False, cpu_only=False, manual=False, manual_window_size=0, final_output_path=None):
def __init__(self, input_data, type, face_type=None, debug_dir=None, multi_gpu=False, cpu_only=False, manual=False, manual_window_size=0, final_output_path=None):
self.input_data = input_data
self.type = type
self.image_size = image_size
self.face_type = face_type
self.debug_dir = debug_dir
self.final_output_path = final_output_path
@ -361,7 +363,6 @@ class ExtractSubprocessor(Subprocessor):
#override
def process_info_generator(self):
base_dict = {'type' : self.type,
'image_size': self.image_size,
'face_type': self.face_type,
'debug_dir': self.debug_dir,
'final_output_dir': str(self.final_output_path),
@ -560,6 +561,7 @@ class ExtractSubprocessor(Subprocessor):
self.landmarks = landmarks[0]
(h,w,c) = self.image.shape
self.image_size = h
if not self.hide_help:
image = cv2.addWeighted (self.image,1.0,self.text_lines_img,1.0,0)
@ -731,7 +733,6 @@ def extract_fanseg(input_dir, device_args={} ):
data = ExtractSubprocessor ([ ExtractSubprocessor.Data(filename) for filename in paths_to_extract ], 'fanseg', multi_gpu=multi_gpu, cpu_only=cpu_only).run()
def extract_umd_csv(input_file_csv,
image_size=256,
face_type='full_face',
device_args={} ):
@ -798,7 +799,7 @@ def extract_umd_csv(input_file_csv,
data = ExtractSubprocessor (data, 'landmarks', multi_gpu=multi_gpu, cpu_only=cpu_only).run()
io.log_info ('Performing 3rd pass...')
data = ExtractSubprocessor (data, 'final', image_size, face_type, None, multi_gpu=multi_gpu, cpu_only=cpu_only, manual=False, final_output_path=output_path).run()
data = ExtractSubprocessor (data, 'final', face_type, None, multi_gpu=multi_gpu, cpu_only=cpu_only, manual=False, final_output_path=output_path).run()
faces_detected += sum([d.faces_detected for d in data])
@ -814,7 +815,6 @@ def main(input_dir,
manual_fix=False,
manual_output_debug_fix=False,
manual_window_size=1368,
image_size=256,
face_type='full_face',
device_args={}):
@ -867,16 +867,16 @@ def main(input_dir,
if images_found != 0:
if detector == 'manual':
io.log_info ('Performing manual extract...')
data = ExtractSubprocessor ([ ExtractSubprocessor.Data(filename) for filename in input_path_image_paths ], 'landmarks', image_size, face_type, debug_dir, cpu_only=cpu_only, manual=True, manual_window_size=manual_window_size).run()
data = ExtractSubprocessor ([ ExtractSubprocessor.Data(filename) for filename in input_path_image_paths ], 'landmarks', face_type, debug_dir, cpu_only=cpu_only, manual=True, manual_window_size=manual_window_size).run()
else:
io.log_info ('Performing 1st pass...')
data = ExtractSubprocessor ([ ExtractSubprocessor.Data(filename) for filename in input_path_image_paths ], 'rects-'+detector, image_size, face_type, debug_dir, multi_gpu=multi_gpu, cpu_only=cpu_only, manual=False).run()
data = ExtractSubprocessor ([ ExtractSubprocessor.Data(filename) for filename in input_path_image_paths ], 'rects-'+detector, face_type, debug_dir, multi_gpu=multi_gpu, cpu_only=cpu_only, manual=False).run()
io.log_info ('Performing 2nd pass...')
data = ExtractSubprocessor (data, 'landmarks', image_size, face_type, debug_dir, multi_gpu=multi_gpu, cpu_only=cpu_only, manual=False).run()
data = ExtractSubprocessor (data, 'landmarks', face_type, debug_dir, multi_gpu=multi_gpu, cpu_only=cpu_only, manual=False).run()
io.log_info ('Performing 3rd pass...')
data = ExtractSubprocessor (data, 'final', image_size, face_type, debug_dir, multi_gpu=multi_gpu, cpu_only=cpu_only, manual=False, final_output_path=output_path).run()
data = ExtractSubprocessor (data, 'final', face_type, debug_dir, multi_gpu=multi_gpu, cpu_only=cpu_only, manual=False, final_output_path=output_path).run()
faces_detected += sum([d.faces_detected for d in data])
if manual_fix:
@ -885,8 +885,8 @@ def main(input_dir,
else:
fix_data = [ ExtractSubprocessor.Data(d.filename) for d in data if d.faces_detected == 0 ]
io.log_info ('Performing manual fix for %d images...' % (len(fix_data)) )
fix_data = ExtractSubprocessor (fix_data, 'landmarks', image_size, face_type, debug_dir, manual=True, manual_window_size=manual_window_size).run()
fix_data = ExtractSubprocessor (fix_data, 'final', image_size, face_type, debug_dir, multi_gpu=multi_gpu, cpu_only=cpu_only, manual=False, final_output_path=output_path).run()
fix_data = ExtractSubprocessor (fix_data, 'landmarks', face_type, debug_dir, manual=True, manual_window_size=manual_window_size).run()
fix_data = ExtractSubprocessor (fix_data, 'final', face_type, debug_dir, multi_gpu=multi_gpu, cpu_only=cpu_only, manual=False, final_output_path=output_path).run()
faces_detected += sum([d.faces_detected for d in fix_data])

7
mainscripts/MaskEditorTool.py
View file

@ -25,8 +25,8 @@ class MaskEditor:
self.img = imagelib.normalize_channels (img,3)
h, w, c = img.shape
if h != w and w != 256:
#to support any square res, scale img,mask and ie_polys to 256, then scale ie_polys back on .get_ie_polys()
if h != w or w != 256:
self.img = cv2.resize(img, (256,256))
raise Exception ("MaskEditor does not support image size != 256x256")
ph, pw = h // 4, w // 4 #pad wh
@ -257,7 +257,7 @@ class MaskEditor:
preview_images += [ np.concatenate (prev_images, axis=1) ]
img = np.full ( (prh,prw, sc), (0,0,1), dtype=np.float )
img[border:-border,border:-border] = cv2.resize( self.img, max_wh_bordered )
img[border:-border,border:-border] = cv2.resize( self.img, 256 )
preview_images += [ img ]
@ -276,7 +276,6 @@ class MaskEditor:
self.preview_images = np.clip(preview_images * 255, 0, 255 ).astype(np.uint8)
status_img = self.get_screen_status_block( screens.shape[1], screens.shape[2] )
result = np.concatenate ( [self.preview_images, screens, status_img], axis=0 )
return result