Spaces:
Paused
Paused
File size: 2,737 Bytes
06d31de |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 |
# facial_diagnostics.py
import cv2
import numpy as np
import math
from statistics import median
# Facial metric helpers
LEFT_EYE_IDX = [33, 160, 158, 133, 153, 144]
RIGHT_EYE_IDX = [362, 385, 387, 263, 373, 380]
MOUTH_IDX = [13, 14, 78, 308]
SMILE_LEFT = 61
SMILE_RIGHT = 291
def eye_aspect_ratio(pts, idx):
a = np.linalg.norm(np.array(pts[idx[1]]) - np.array(pts[idx[5]]))
b = np.linalg.norm(np.array(pts[idx[2]]) - np.array(pts[idx[4]]))
c = np.linalg.norm(np.array(pts[idx[0]]) - np.array(idx[3])) + 1e-8
return float((a + b) / (2.0 * c))
def mouth_ratio(pts):
top, bottom, left, right = pts[MOUTH_IDX[0]], pts[MOUTH_IDX[1]], pts[MOUTH_IDX[2]], pts[MOUTH_IDX[3]]
return float(np.linalg.norm(np.array(top) - np.array(bottom)) /
(np.linalg.norm(np.array(left) - np.array(right)) + 1e-8))
def head_tilt_angle(pts):
left = np.mean([pts[33], pts[133]], axis=0)
right = np.mean([pts[362], pts[263]], axis=0)
diff = np.array(right) - np.array(left)
return float(math.degrees(math.atan2(diff[1], diff[0])))
def smile_symmetry(pts):
left = np.array(pts[SMILE_LEFT])
right = np.array(pts[SMILE_RIGHT])
center = (left + right) / 2
L = np.linalg.norm(left - center)
R = np.linalg.norm(right - center)
if L + R == 0:
return 1.0
return float(min(L, R) / max(L, R))
# Face mesh
import mediapipe as mp
mp_face_mesh = mp.solutions.face_mesh
FACE_MESH = mp_face_mesh.FaceMesh(
static_image_mode=False,
max_num_faces=1,
refine_landmarks=True,
min_detection_confidence=0.6,
min_tracking_confidence=0.6
)
def analyze_frame(frame):
h, w, _ = frame.shape
rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
res = FACE_MESH.process(rgb)
if not res.multi_face_landmarks:
return {"found": False}
lm = res.multi_face_landmarks[0]
pts = [(int(p.x * w), int(p.y * h)) for p in lm.landmark]
ear = (eye_aspect_ratio(pts, LEFT_EYE_IDX) + eye_aspect_ratio(pts, RIGHT_EYE_IDX)) / 2
mratio = mouth_ratio(pts)
tilt = head_tilt_angle(pts)
symmetry = smile_symmetry(pts)
return {
"found": True,
"ear": float(ear),
"mouth_ratio": float(mratio),
"head_tilt": float(tilt),
"smile_sym": float(symmetry)
}
def aggregate_metrics(metrics):
if not metrics:
return {}
ears = [m["ear"] for m in metrics]
mouths = [m["mouth_ratio"] for m in metrics]
tilts = [m["head_tilt"] for m in metrics]
smiles = [m["smile_sym"] for m in metrics]
return {
"count": len(metrics),
"ear_median": median(ears),
"mouth_median": median(mouths),
"tilt_median": median(tilts),
"smile_median": median(smiles)
}
|