fix: add support for customizable score and NMS thresholds

frigate/detectors/plugins/deepx.py

@@ -1136,7 +1136,13 @@ class DeepxDetector(DetectionApi):
                 self.nms_threshold,
             )
 
-        return post_process_yolo(outputs, self.width, self.height)
+        return post_process_yolo(
+            outputs,
+            self.width,
+            self.height,
+            self.score_threshold,
+            self.nms_threshold,
+        )
 
     def detect_raw(self, tensor_input):
         """Run inference and decode the DEEPX model output."""

frigate/test/test_post_processing.py

@@ -81,9 +81,21 @@ class TestYoloNmsPostProcess(unittest.TestCase):
         self.assertEqual(len(detections), 1)
         np.testing.assert_allclose(detections[0], [0, 0.90, *A_ROW], atol=2e-3)
 
+    def test_honors_caller_supplied_thresholds(self):
+        """A detector with its own thresholds must not be held to the 0.4
+        default, which would drop this 0.3 confidence box."""
+        output = self._single_output([[A_CX, A_CY, A_W, A_H, 0.30, 0.0]])
+
+        self.assertEqual(len(kept(post_process_yolo(output, WIDTH, HEIGHT))), 0)
+
+        detections = kept(post_process_yolo(output, WIDTH, HEIGHT, 0.25, 0.45))
+
+        self.assertEqual(len(detections), 1)
+        np.testing.assert_allclose(detections[0], [0, 0.30, *A_ROW], atol=2e-3)
+
 
 class TestMultipartYoloPostProcess(unittest.TestCase):
-    def _multipart_output(self) -> list[np.ndarray]:
+    def _multipart_output(self, a_conf=0.95, b_conf=0.90) -> list[np.ndarray]:
         """Build a 3-scale anchor-based YOLO output containing boxes A and B,
         both decoded through anchor 0 of the stride-32 scale."""
         outputs = [
@@ -94,8 +106,8 @@ class TestMultipartYoloPostProcess(unittest.TestCase):
         stride, (anchor_w, anchor_h) = 32, (142, 110)
 
         for cx, cy, w, h, conf, class_channel in [
-            (A_CX, A_CY, A_W, A_H, 0.95, 5),  # class 0
-            (B_CX, B_CY, B_W, B_H, 0.90, 6),  # class 1
+            (A_CX, A_CY, A_W, A_H, a_conf, 5),  # class 0
+            (B_CX, B_CY, B_W, B_H, b_conf, 6),  # class 1
         ]:
             cell_x, cell_y = int(cx // stride), int(cy // stride)
             dx = (cx / stride - cell_x + 0.5) / 2
@@ -127,6 +139,19 @@ class TestMultipartYoloPostProcess(unittest.TestCase):
 
         self.assertEqual(len(detections), 0)
 
+    def test_honors_caller_supplied_thresholds(self):
+        """Same as the single-tensor case: the multi-part branch must filter
+        at the caller's threshold, not the 0.4 default."""
+        outputs = self._multipart_output(a_conf=0.35, b_conf=0.30)
+
+        self.assertEqual(len(kept(post_process_yolo(outputs, WIDTH, HEIGHT))), 0)
+
+        detections = kept(post_process_yolo(outputs, WIDTH, HEIGHT, 0.25, 0.45))
+
+        self.assertEqual(len(detections), 2)
+        np.testing.assert_allclose(detections[0], [0, 0.35, *A_ROW], atol=2e-3)
+        np.testing.assert_allclose(detections[1], [1, 0.30, *B_ROW], atol=2e-3)
+
 
 class TestYoloxPostProcess(unittest.TestCase):
     def test_keeps_separated_objects_far_from_origin(self):

frigate/util/model.py

@@ -134,6 +134,8 @@ def __post_process_multipart_yolo(
     output_list,
     width,
     height,
+    score_threshold=0.4,
+    nms_threshold=0.4,
 ):
     anchors = [
         [(12, 16), (19, 36), (40, 28)],
@@ -166,7 +168,7 @@ def __post_process_multipart_yolo(
                     class_conf = class_probs[class_id]
                     conf = class_conf * pred[4]
 
-                    if conf < 0.4:
+                    if conf < score_threshold:
                         continue
 
                     dx = pred[0]
@@ -191,8 +193,8 @@ def __post_process_multipart_yolo(
     indices = cv2.dnn.NMSBoxes(
         bboxes=xyxy_to_xywh_for_nms(all_boxes),
         scores=all_scores,
-        score_threshold=0.4,
-        nms_threshold=0.4,
+        score_threshold=score_threshold,
+        nms_threshold=nms_threshold,
     )
 
     results = np.zeros((20, 6), np.float32)
@@ -214,7 +216,13 @@ def __post_process_multipart_yolo(
     return results
 
 
-def __post_process_nms_yolo(predictions: np.ndarray, width, height) -> np.ndarray:
+def __post_process_nms_yolo(
+    predictions: np.ndarray,
+    width,
+    height,
+    score_threshold=0.4,
+    nms_threshold=0.4,
+) -> np.ndarray:
     predictions = np.squeeze(predictions)
 
     # transpose the output so it has order (inferences, class_ids)
@@ -222,8 +230,8 @@ def __post_process_nms_yolo(predictions: np.ndarray, width, height) -> np.ndarra
         predictions = predictions.T
 
     scores = np.max(predictions[:, 4:], axis=1)
-    predictions = predictions[scores > 0.4, :]
-    scores = scores[scores > 0.4]
+    predictions = predictions[scores > score_threshold, :]
+    scores = scores[scores > score_threshold]
     class_ids = np.argmax(predictions[:, 4:], axis=1)
 
     # Rescale box
@@ -237,7 +245,10 @@ def __post_process_nms_yolo(predictions: np.ndarray, width, height) -> np.ndarra
 
     # run NMS
     indices = cv2.dnn.NMSBoxes(
-        xyxy_to_xywh_for_nms(boxes), scores, score_threshold=0.4, nms_threshold=0.4
+        xyxy_to_xywh_for_nms(boxes),
+        scores,
+        score_threshold=score_threshold,
+        nms_threshold=nms_threshold,
     )
     detections = np.zeros((20, 6), np.float32)
     for i, (bbox, confidence, class_id) in enumerate(
@@ -258,11 +269,21 @@ def __post_process_nms_yolo(predictions: np.ndarray, width, height) -> np.ndarra
     return detections
 
 
-def post_process_yolo(output: list[np.ndarray], width: int, height: int) -> np.ndarray:
+def post_process_yolo(
+    output: list[np.ndarray],
+    width: int,
+    height: int,
+    score_threshold: float = 0.4,
+    nms_threshold: float = 0.4,
+) -> np.ndarray:
     if len(output) > 1:
-        return __post_process_multipart_yolo(output, width, height)
+        return __post_process_multipart_yolo(
+            output, width, height, score_threshold, nms_threshold
+        )
     else:
-        return __post_process_nms_yolo(output[0], width, height)
+        return __post_process_nms_yolo(
+            output[0], width, height, score_threshold, nms_threshold
+        )
 
 
 def post_process_yolox(