From 7fc0210ce61032eb777ee80bb6f7c7bb251c8ade Mon Sep 17 00:00:00 2001
From: xsl <youmiss@163.com>
Date: Tue, 21 Jul 2026 16:46:05 +0800
Subject: [PATCH] =?UTF-8?q?fix(pose):=20=E4=BF=AE=E6=AD=A3=E6=AD=A3?=
 =?UTF-8?q?=E9=9D=A2=E7=85=A7=E8=A2=AB=E8=AF=AF=E5=88=A4=E4=B8=BA1003?=
 =?UTF-8?q?=EF=BC=88solvePnP=E7=BF=BB=E8=BD=AC=E8=A7=A3=EF=BC=89?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

ITERATIVE 偶发收敛到相机后方(tz<0)，roll≈±180° 超阈值，
把正面照误判为非正面。检测到负深度时回退 SQPNP 重解正深度解。
补充回归测试。

Co-authored-by: Cursor <cursoragent@cursor.com>
---
 face_analysis/pose.py | 11 ++++++++++-
 tests/test_pose.py    | 38 ++++++++++++++++++++++++++++++++++++++
 2 files changed, 48 insertions(+), 1 deletion(-)

diff --git a/face_analysis/pose.py b/face_analysis/pose.py
index 423af00..57d385c 100644
--- a/face_analysis/pose.py
+++ b/face_analysis/pose.py
@@ -50,12 +50,21 @@ def estimate_head_pose(landmarks, image_width, image_height):
                            [0, 0, 1]], dtype=np.float64)
     dist = np.zeros((4, 1))  # 假设无畸变
 
-    success, rvec, _tvec = cv2.solvePnP(
+    success, rvec, tvec = cv2.solvePnP(
         _MODEL_POINTS, image_points, cam_matrix, dist,
         flags=cv2.SOLVEPNP_ITERATIVE,
     )
     if not success:
         return None
+    # ITERATIVE 偶发收敛到相机后方的翻转解（tz<0），此时 roll 落在 ±180° 附近，
+    # 会把真正的正面照误判为 1003。改用 SQPNP 重解正深度解。
+    if float(tvec[2, 0]) < 0:
+        ok2, rvec2, tvec2 = cv2.solvePnP(
+            _MODEL_POINTS, image_points, cam_matrix, dist,
+            flags=cv2.SOLVEPNP_SQPNP,
+        )
+        if ok2 and float(tvec2[2, 0]) > 0:
+            rvec = rvec2
     rot, _ = cv2.Rodrigues(rvec)
     # 在「相机坐标系」(x右 y下 z内) 下抽取 Tait-Bryan 欧拉角，物理含义对齐：
     #   yaw   = 绕 Y（竖轴）转 → 左右扭头
diff --git a/tests/test_pose.py b/tests/test_pose.py
index 11b5695..663ac9e 100644
--- a/tests/test_pose.py
+++ b/tests/test_pose.py
@@ -62,3 +62,41 @@ def test_threshold_gating_rejects_when_zeroed():
 def test_pose_none_is_not_blocked():
     """solvePnP 失败（返回 None）时不拦截，check_frontal_face 返回 True。"""
     assert pose.estimate_head_pose.__doc__  # 占位，确保导入
+
+
+def test_iterative_flipped_solution_falls_back_to_sqpnp():
+    """回归：部分正面照上 ITERATIVE 会解出 tz<0、roll≈±180°，应回退 SQPNP。
+
+    像素点取自一张真实正面短发照（720×945）；裸跑 ITERATIVE 会得到负深度。
+    """
+    W, H = 720, 945
+    # 鼻尖 / 下巴 / 左眼外 / 右眼外 / 左嘴角 / 右嘴角（像素）
+    px = [
+        (358.32715988, 600.98652095),
+        (346.83344364, 779.63507116),
+        (242.94779778, 466.76155195),
+        (478.43703747, 480.10321766),
+        (284.13277388, 679.95527387),
+        (422.19510555, 684.31899190),
+    ]
+    lm = [_LM(0.5, 0.5) for _ in range(478)]
+    for idx, (u, v) in zip(PNP_INDICES, px):
+        lm[idx] = _LM(u / W, v / H)
+
+    class _Holder:
+        landmark = lm
+
+    holder = _Holder()
+    # 确认裸 ITERATIVE 确实是翻转解（否则本回归失去意义）
+    image_points = np.array(px, dtype=np.float64)
+    cam = np.array([[float(W), 0, W / 2], [0, float(W), H / 2], [0, 0, 1]],
+                   dtype=np.float64)
+    ok, rvec, tvec = cv2.solvePnP(
+        _MODEL_POINTS, image_points, cam, np.zeros((4, 1)),
+        flags=cv2.SOLVEPNP_ITERATIVE,
+    )
+    assert ok and float(tvec[2, 0]) < 0
+
+    yaw, pitch, roll = estimate_head_pose(holder, W, H)
+    assert abs(roll) < 30, f"roll 应被纠正，实际 roll={roll}"
+    assert check_frontal_face(holder, W, H) is True