Update emotion.py

emotion.py

@@ -68,7 +68,7 @@ SCORE_CHOICES = [SCORE_LABELS[i] for i in [1, 2, 3, 4, 5]]
 # =========================
 @dataclass
 class AppState:
-    mode: str = "攝影機辨識" 
+    mode: str = "攝影機辨識"
 
     running: bool = True
     finished: bool = False
@@ -90,14 +90,17 @@ class AppState:
     final_emo: Optional[str] = None
     final_conf: Optional[float] = None
 
-    final_score: int = 3  
+    final_score: int = 3
+
+    # ✅ 新增：最近有沒有臉（避免節流期間一直紅框）
+    last_face_t: float = 0.0
+    face_in_box: bool = False
 
 
 # =========================
 # UI HTML (強制黑色文字)
 # =========================
 def _hint_html(msg: str) -> str:
-    # 加入 color:#000000 強制黑色
     return f"""
     <div style="border-radius:14px;padding:14px;border:1px dashed #DADADA;background:#FAFAFA;color:#000000;">
       {msg}
@@ -126,7 +129,6 @@ def _result_card_html(emo_key: str, conf: Optional[float]) -> str:
     conf_txt = "" if conf is None else f"{conf:.1f}%"
     score_desc = SCORE_LABELS[score].split("（")[1].rstrip("）")
 
-    # ✅ 強制設定 color: #000000
     return f"""
     <div style="
         border-radius:16px;
@@ -175,25 +177,16 @@ def _downsample_rgb(frame: np.ndarray, target_w: int) -> np.ndarray:
     img = img.resize((target_w, new_h))
     return np.array(img)
 
+
 def _zoom_center(frame: np.ndarray, zoom: float) -> np.ndarray:
-    """
-    裁切畫面中心以模擬數位變焦。
-    zoom: 放大倍率 (例如 1.5)
-    """
     if frame is None or zoom <= 1.0:
         return frame
-        
+
     h, w = frame.shape[:2]
-    
-    # 計算新的寬高
     new_w = int(w / zoom)
     new_h = int(h / zoom)
-    
-    # 計算起始點 (讓裁切框置中)
     start_x = (w - new_w) // 2
     start_y = (h - new_h) // 2
-    
-    # 進行裁切
     cropped = frame[start_y : start_y + new_h, start_x : start_x + new_w]
     return cropped
 
@@ -201,12 +194,11 @@ def _zoom_center(frame: np.ndarray, zoom: float) -> np.ndarray:
 # =========================
 # 引導框（臉要放哪裡）
 # =========================
-GUIDE_W_RATIO = 0.55   # 框寬佔畫面比例（可調）
-GUIDE_H_RATIO = 0.70   # 框高佔畫面比例（可調）
-GUIDE_Y_OFFSET = 0.00  # 框中心往上(-)或往下(+)位移比例（可調，例如 -0.05）
+GUIDE_W_RATIO = 0.55
+GUIDE_H_RATIO = 0.70
+GUIDE_Y_OFFSET = 0.00
 
 def _guide_box(frame: np.ndarray) -> Tuple[int, int, int, int]:
-    """回傳引導框 (x1,y1,x2,y2)"""
     h, w = frame.shape[:2]
     gw = int(w * GUIDE_W_RATIO)
     gh = int(h * GUIDE_H_RATIO)
@@ -221,7 +213,6 @@ def _guide_box(frame: np.ndarray) -> Tuple[int, int, int, int]:
     return x1, y1, x2, y2
 
 def _bbox_center_in_box(bbox_xywh: Tuple[float, float, float, float], box_xyxy: Tuple[int, int, int, int]) -> bool:
-    """判斷臉框中心是否落在引導框內"""
     x, y, w, h = bbox_xywh
     cx = x + w / 2.0
     cy = y + h / 2.0
@@ -229,9 +220,6 @@ def _bbox_center_in_box(bbox_xywh: Tuple[float, float, float, float], box_xyxy:
     return (x1 <= cx <= x2) and (y1 <= cy <= y2)
 
 def _draw_guide_overlay(frame_rgb: np.ndarray, detected_in_box: bool) -> np.ndarray:
-    """
-    在畫面上畫引導框：未偵測到(或不在框內)=紅框；偵測到且在框內=綠框
-    """
     if frame_rgb is None:
         return frame_rgb
 
@@ -240,27 +228,27 @@ def _draw_guide_overlay(frame_rgb: np.ndarray, detected_in_box: bool) -> np.ndar
 
     x1, y1, x2, y2 = _guide_box(frame_rgb)
 
-    # 顏色：紅 / 綠（RGB）
+    # 紅/綠框
     color = (0, 200, 0) if detected_in_box else (220, 0, 0)
 
-    # 線寬（跟解析度稍微連動）
     h, w = frame_rgb.shape[:2]
     thickness = max(3, int(min(w, h) * 0.008))
 
-    # 畫框（加粗）
     for t in range(thickness):
         draw.rectangle([x1 - t, y1 - t, x2 + t, y2 + t], outline=color)
 
-    # 提示文字
-    msg = "偵測中…" if detected_in_box else "請把臉放進框內"
+    msg = "已偵測到人臉，開始辨識情緒中…" if detected_in_box else "未偵測到人臉，請把臉放進框內"
     tx = x1
     ty = max(0, y1 - int(thickness * 6))
-    draw.rectangle([tx, ty, tx + 240, ty + 32], fill=(255, 255, 255))
+    draw.rectangle([tx, ty, tx + 360, ty + 32], fill=(255, 255, 255))
     draw.text((tx + 6, ty + 6), msg, fill=(0, 0, 0))
 
     return np.array(img)
 
 
+# =========================
+# Tracking / IoU
+# =========================
 def _iou(a: Tuple[int, int, int, int], b: Tuple[int, int, int, int]) -> float:
     ax, ay, aw, ah = a
     bx, by, bw, bh = b
@@ -325,6 +313,10 @@ def _reset_state_for_camera(st: AppState) -> AppState:
     st.final_emo = None
     st.final_conf = None
     st.final_score = 3
+
+    # ✅ 新增欄位也要 reset
+    st.last_face_t = 0.0
+    st.face_in_box = False
     return st
 
 
@@ -362,20 +354,25 @@ def on_stream(frame_rgb: np.ndarray, st: AppState):
             st,
             gr.update(value="重新攝影機辨識"),
         )
-    
-    frame_rgb = _zoom_center(frame_rgb, ZOOM_FACTOR)
 
-    # 預設：沒偵測��（或不在引導框內）-> 紅框
-    detected_in_box = False
+    frame_rgb = _zoom_center(frame_rgb, ZOOM_FACTOR)
 
     now = time.time()
+
+    # ✅ 節流時沿用最近一次「臉是否在框內」狀態（避免一直紅框）
+    detected_in_box_recent = st.face_in_box and (now - st.last_face_t) < 1.0
+
     if now - st.last_analyze_t < ANALYZE_EVERY_SEC:
-        show = st.stable_emo or "unknown"
-        annotated = _draw_guide_overlay(frame_rgb, detected_in_box)
+        annotated = _draw_guide_overlay(frame_rgb, detected_in_box_recent)
+        if detected_in_box_recent:
+            result_html = _hint_html("已偵測到人臉，開始辨識情緒中…")
+        else:
+            result_html = _hint_html("未偵測到人臉，請把臉放進框內")
+
         return (
             gr.update(visible=True, value=annotated),
-            gr.update(value=_result_card_html(show, st.final_conf)),
-            gr.update(value=_score_radio_value(EMO_TO_SCORE.get(show, 3))),
+            gr.update(value=result_html),
+            gr.update(value=_score_radio_value(3)),
             gr.update(value=_camera_hint_before_start()),
             st,
             gr.update(value="重新攝影機辨識"),
@@ -391,20 +388,27 @@ def on_stream(frame_rgb: np.ndarray, st: AppState):
         enforce_detection=False,
         align=ALIGN_FACE,
     )
+
     pick, bbox = _pick_face_by_tracking(faces, st.track_bbox)
+
+    # ✅ (A) 完全沒臉：紅框 + 不顯示情緒（避免「沒臉也顯示害怕」）
     if pick is None or bbox is None or pick.get("face") is None:
-        show = st.stable_emo or "unknown"
-        annotated = _draw_guide_overlay(frame_rgb, detected_in_box)
+        st.track_bbox = None
+        st.face_in_box = False
+        st.last_face_t = now
+        st.stable_emo = "unknown"
+
+        annotated = _draw_guide_overlay(frame_rgb, False)
         return (
             gr.update(visible=True, value=annotated),
-            gr.update(value=_result_card_html(show, st.final_conf)),
-            gr.update(value=_score_radio_value(EMO_TO_SCORE.get(show, 3))),
+            gr.update(value=_hint_html("未偵測到人臉，請把臉放進框內")),
+            gr.update(value=_score_radio_value(3)),
             gr.update(value=_camera_hint_before_start()),
             st,
             gr.update(value="重新攝影機辨識"),
         )
 
-    # bbox 是 small 座標，換算回 frame_rgb 座標後判斷是否在引導框內
+    # bbox 是 small 座標，換算回 frame_rgb
     sh, sw = small.shape[:2]
     fh, fw = frame_rgb.shape[:2]
     sx = fw / max(1, sw)
@@ -412,10 +416,28 @@ def on_stream(frame_rgb: np.ndarray, st: AppState):
 
     x, y, w, h = bbox
     bbox_full = (x * sx, y * sy, w * sx, h * sy)
+
     guide = _guide_box(frame_rgb)
     detected_in_box = _bbox_center_in_box(bbox_full, guide)
 
+    # ✅ 更新「臉狀態」
+    st.last_face_t = now
+    st.face_in_box = detected_in_box
     st.track_bbox = bbox
+
+    # ✅ (B) 有臉但不在框內：紅框，不做情緒辨識
+    if not detected_in_box:
+        annotated = _draw_guide_overlay(frame_rgb, False)
+        return (
+            gr.update(visible=True, value=annotated),
+            gr.update(value=_hint_html("已偵測到人臉，但請把臉放進框內")),
+            gr.update(value=_score_radio_value(3)),
+            gr.update(value=_camera_hint_before_start()),
+            st,
+            gr.update(value="重新攝影機辨識"),
+        )
+
+    # ✅ (C) 有臉且在框內：綠框，才開始 analyze
     face = pick["face"]
     if face.dtype != np.uint8:
         face = np.clip(face * 255.0, 0, 255).astype(np.uint8)
@@ -429,13 +451,13 @@ def on_stream(frame_rgb: np.ndarray, st: AppState):
         result = result[0]
     emo_dict = result.get("emotion", None)
 
+    # 若分析結果不完整：仍顯示綠框 + 辨識中提示
     if not isinstance(emo_dict, dict) or len(emo_dict) == 0:
-        show = st.stable_emo or "unknown"
-        annotated = _draw_guide_overlay(frame_rgb, detected_in_box)
+        annotated = _draw_guide_overlay(frame_rgb, True)
         return (
             gr.update(visible=True, value=annotated),
-            gr.update(value=_result_card_html(show, st.final_conf)),
-            gr.update(value=_score_radio_value(EMO_TO_SCORE.get(show, 3))),
+            gr.update(value=_hint_html("已偵測到人臉，開始辨識情緒中…")),
+            gr.update(value=_score_radio_value(3)),
             gr.update(value=_camera_hint_before_start()),
             st,
             gr.update(value="重新攝影機辨識"),
@@ -444,13 +466,13 @@ def on_stream(frame_rgb: np.ndarray, st: AppState):
     top_key = max(emo_dict, key=lambda k: float(emo_dict[k]))
     top_conf = float(emo_dict[top_key])
 
+    # 信心不足：仍綠框 + 辨識中
     if top_conf < CONF_TH:
-        show = st.stable_emo or "unknown"
-        annotated = _draw_guide_overlay(frame_rgb, detected_in_box)
+        annotated = _draw_guide_overlay(frame_rgb, True)
         return (
             gr.update(visible=True, value=annotated),
-            gr.update(value=_result_card_html(show, st.final_conf)),
-            gr.update(value=_score_radio_value(EMO_TO_SCORE.get(show, 3))),
+            gr.update(value=_hint_html("已偵測到人臉，開始辨識情緒中…")),
+            gr.update(value=_score_radio_value(3)),
             gr.update(value=_camera_hint_before_start()),
             st,
             gr.update(value="重新攝影機辨識"),
@@ -499,7 +521,7 @@ def on_stream(frame_rgb: np.ndarray, st: AppState):
             gr.update(value="重新攝影機辨識"),
         )
 
-    annotated = _draw_guide_overlay(frame_rgb, detected_in_box)
+    annotated = _draw_guide_overlay(frame_rgb, True)
     return (
         gr.update(visible=True, value=annotated),
         gr.update(value=_result_card_html(stable_key, top_conf)),
@@ -513,16 +535,15 @@ def on_stream(frame_rgb: np.ndarray, st: AppState):
 # =========================
 # Mode / Buttons
 # =========================
-
 def on_restart(st: AppState):
     st = _reset_state_for_camera(st)
     return (
-        gr.update(visible=True),                               
-        gr.update(value=_hint_html("等待影像…（尚未開始串流）")),  
-        gr.update(value=_score_radio_value(3)),                
-        gr.update(value=_camera_hint_before_start()),          
+        gr.update(visible=True),
+        gr.update(value=_hint_html("等待影像…（尚未開始串流）")),
+        gr.update(value=_score_radio_value(3)),
+        gr.update(value=_camera_hint_before_start()),
         st,
-        gr.update(value="重新攝影機辨識"),                      
+        gr.update(value="重新攝影機辨識"),
     )
 
 
@@ -552,13 +573,12 @@ def on_stop(st: AppState):
 # =========================
 # Gradio UI
 # =========================
-
 if __name__ == "__main__":
     with gr.Blocks(title="Emotion Detector", css=css) as demo:
         st = gr.State(AppState())
 
         gr.Markdown("## 情緒辨識")
-        
+
         hint = gr.HTML(_camera_hint_before_start())
 
         cam = gr.Image(