Commit 2abe769a authored by Olegs NIKISINS's avatar Olegs NIKISINS

Updated BATL HLDI: added rc, skin/non-skin annotations, multi-channel laoding functionalities

parent c7e0f345
Pipeline #23475 passed with stage
in 16 minutes and 52 seconds
This diff is collapsed.
Files contain annotations for the color images in the training set of the Idiap BATL database.
A region correponding to non-skin pixels is annotated in the "Attack" samples, and
region corresponding to skin pixels is annotated in the "Bona-fide" samples.
Annotations define a top-left and bottom-right corners of the ROI.
Note, annotations are made in images, which are first preprocessed. In this case
preprocessing is face detection, cropping and alignment.
The following pre-preprocessor was used before annotating:
FACE_SIZE = 128 # The size of the resulting face
RGB_OUTPUT_FLAG = True # RGB output
USE_FACE_ALIGNMENT = True # use annotations, which are coming from the database
MAX_IMAGE_SIZE = None # no limiting here
FACE_DETECTION_METHOD = None # use annotations
MIN_FACE_SIZE = 50 # skip small faces
_image_preprocessor = FaceCropAlign(face_size = FACE_SIZE,
rgb_output_flag = RGB_OUTPUT_FLAG,
use_face_alignment = USE_FACE_ALIGNMENT,
max_image_size = MAX_IMAGE_SIZE,
face_detection_method = FACE_DETECTION_METHOD,
min_face_size = MIN_FACE_SIZE)
_frame_selector = FrameSelector(selection_style = "all")
_preprocessor_rgb = Wrapper(preprocessor = _image_preprocessor,
frame_selector = _frame_selector)
The following setting were used for the database instance:
database = BatlPadDatabase(
protocol='grandtest-color*infrared*depth-1', # just one frame is selected for annotation
annotations_temp_dir="", # don't use precomputed annotations
landmark_detect_method="mtcnn", # mtcnn is used for landmark detection, above preprocessor is using these landmarks (FACE_DETECTION_METHOD = None)
exclude_attacks_list=['makeup'], # makeup is excluded from annotations
Note: eye-regions and medical glasses are also considered as BF patches in bona-fide faces.
Markdown is supported
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment