Update app.py

app.py

@@ -11,44 +11,43 @@ API_URL = f"https://serverless.roboflow.com/{MODEL_ID}"
 
 
 # -----------------------------------
-# 2) confidence → color 규칙
+# 2) confidence 색상 규칙
 # -----------------------------------
 def get_color(conf):
     if conf < 0.30:
-        return (0, 255, 0, 120)     # Green
+        return (0, 255, 0, 120)      # Green
     elif conf < 0.60:
-        return (255, 165, 0, 120)   # Orange
+        return (255, 165, 0, 120)    # Orange
     else:
-        return (255, 0, 0, 120)     # Red
+        return (255, 0, 0, 120)      # Red
 
 
 # -----------------------------------
-# 3) Roboflow 호출 + mask 시각화 + JSON 출력
+# 3) Roboflow 호출 + 다중 segmentation 시각화
 # -----------------------------------
 def predict(image: Image.Image):
     temp_path = "temp.jpg"
     image.save(temp_path)
 
-    # Roboflow API 호출
+    # threshold 낮추기 → crack 감지 활성화
     with open(temp_path, "rb") as f:
         response = requests.post(
             API_URL,
             files={"file": f},
-                data={
-                    "api_key": API_KEY,
-                    "confidence": 0.2  # <-- 임계값 낮추기
-                }
+            data={
+                "api_key": API_KEY,
+                "confidence": 0.15   # ★ 매우 중요: threshold 낮추기
+            }
         )
 
     result = response.json()
     preds = result.get("predictions", [])
 
-    # 베이스 이미지
     base = image.convert("RGBA")
-    overlay = Image.new("RGBA", base.size, (0, 0, 0, 0))
+    overlay = Image.new("RGBA", base.size, (0,0,0,0))
     draw = ImageDraw.Draw(overlay, "RGBA")
 
-    # 예측 없음 → 정상 처리
+    # 객체 없음 → 정상 처리
     if not preds:
         blended = Image.alpha_composite(base, overlay)
         return blended, {
@@ -58,30 +57,32 @@ def predict(image: Image.Image):
             }]
         }
 
-    # 가장 confidence 높은 객체 선택
-    best = max(preds, key=lambda p: p.get("confidence", 0))
-
-    conf = float(best.get("confidence", 0))
-    label = best.get("class", "crack")
-
-    # segmentation polygon 색칠
-    points = best.get("points")
-    if points:
-        polygon = [(p["x"], p["y"]) for p in points]
-        draw.polygon(polygon, fill=get_color(conf))
+    # ① 다중 crack 시각화 (for all preds)
+    for p in preds:
+        conf = float(p.get("confidence", 0))
+        color = get_color(conf)
+        points = p.get("points")
+        if points:
+            polygon = [(pt["x"], pt["y"]) for pt in points]
+            draw.polygon(polygon, fill=color)
 
     blended = Image.alpha_composite(base, overlay)
 
+    # ② 가장 confidence 높은 crack 하나만 프론트로 전달
+    best = max(preds, key=lambda p: p.get("confidence", 0))
+    best_conf = float(best.get("confidence", 0))
+    best_label = best.get("class", "crack")
+
     return blended, {
         "data": [{
-            "label": label,
-            "confidence": conf
+            "label": best_label,
+            "confidence": best_conf
         }]
     }
 
 
 # -----------------------------------
-# 4) Gradio Interface
+# 4) Gradio UI
 # -----------------------------------
 demo = gr.Interface(
     fn=predict,
@@ -90,8 +91,8 @@ demo = gr.Interface(
         gr.Image(label="Crack Heatmap"),
         gr.JSON(label="Detection Data"),
     ],
-    title="Crack Detection + Heatmap (Roboflow)",
-    description="Segmentation 기반 균열 영역을 색으로 표시하고 JSON을 반환하는 API",
+    title="Crack Detection + Multi-Seg Heatmap",
+    description="여러 균열을 각각 다른 색으로 표시하고, 가장 높은 균열 확률을 UI에 제공합니다.",
     flagging_mode="never"
 )