Improvements to handling

frigate/data_processing/real_time/bird_processor.py

@@ -3,10 +3,12 @@
 import logging
 import os
 
+import cv2
 import numpy as np
 
 from frigate.config import FrigateConfig
 from frigate.const import MODEL_CACHE_DIR
+from frigate.util.object import calculate_region
 
 from ..types import DataProcessorMetrics
 from .api import RealTimeProcessorApi
@@ -75,16 +77,36 @@ class BirdProcessor(RealTimeProcessorApi):
         if obj_data["label"] != "bird":
             return
 
-        self.interpreter.set_tensor(self.tensor_input_details[0]["index"], frame)
+        x, y, x2, y2 = calculate_region(
+            frame.shape,
+            obj_data["box"][0],
+            obj_data["box"][1],
+            obj_data["box"][2],
+            obj_data["box"][3],
+            224,
+            1.4,
+        )
+
+        rgb = cv2.cvtColor(frame, cv2.COLOR_YUV2RGB_I420)
+        input = rgb[
+            y:y2,
+            x:x2,
+        ]
+
+        logger.info(f"input shape is {input.shape}")
+        cv2.imwrite("/media/frigate/test_class.png", input)
+
+        input = np.expand_dims(input, axis=0)
+
+        self.interpreter.set_tensor(self.tensor_input_details[0]["index"], input)
         self.interpreter.invoke()
-        res = self.interpreter.get_tensor(self.tensor_output_details[0]["index"])[0]
-        non_zero_indices = res > 0
-        class_ids = np.argpartition(-res, 20)[:20]
-        class_ids = class_ids[np.argsort(-res[class_ids])]
-        class_ids = class_ids[non_zero_indices[class_ids]]
-        scores = res[class_ids]
-        boxes = np.full((scores.shape[0], 4), -1, np.float32)
-        count = len(scores)
+        res: np.ndarray = self.interpreter.get_tensor(self.tensor_output_details[0]["index"])[0]
+        probs = res / res.sum(axis=0)
+        best_id = np.argmax(probs)
+        score = probs[best_id]
 
     def handle_request(self, request_data):
         return None
+
+    def expire_object(self, object_id):
+        pass