from pyVHR.extraction.utils import MagicLandmarks
from pyVHR.BVP.methods import *
from pyVHR.BVP.filters import *
from pyVHR.extraction.utils import MagicLandmarks
[docs]class Params:
# DEFAULT PARAMETERS
videoFileName = ''
fps_fixed = None
tot_sec = 0
winSize = 6
stride = 1
cuda = False
skin_extractor = 'convexhull' # or faceparsing
approach = 'patches' # or holistic
patches = 'squares' # or rects
type = 'mean'
landmarks_list = MagicLandmarks.equispaced_facial_points
squares_dim = 30.0
rects_dims = []
# extractor params
skin_color_low_threshold = 75
skin_color_high_threshold = 230
sig_color_low_threshold = 75
sig_color_high_threshold = 230
# rgb_filter_th params
color_low_threshold = 75
color_high_threshold = 230
# visualize skin and patches
visualize_skin = True
visualize_patches = True
visualize_landmarks = True
visualize_landmarks_number = True
font_size = 0.3
font_color = (255, 0, 0, 255)
# Pre filtering
# dictionary of {filter_func, params}
pre_filter = [{'filter_func': BPfilter, 'params': {
'minHz': 0.7, 'maxHz': 3.0, 'fps': 'adaptive', 'order': 6}}]
# BVP method
# dictionary of {method_func, device_type, params}
method = {'method_func': cpu_CHROM,
'device_type': 'cpu', 'params': {}}
# Post filtering
# dictionary of {filter_func, params}
post_filter = [{'filter_func': BPfilter, 'params': {
'minHz': 0.7, 'maxHz': 3.0, 'fps': 'adaptive', 'order': 6}}]
# BPM params
minHz = 0.7
maxHz = 4.0
# WELCH: CPU, GPU
# PSD_CLUSTERING: CPU, GPU
# USE psd_clustering only with patches!
BPM_extraction_type = "welch" # or 'psd_clustering'
# Utils
fake_delay = False
resize = True
out_path = None