predict: sanitzation

plugins/coreml/src/coreml/__init__.py

@@ -250,13 +250,13 @@ class CoreMLPlugin(PredictPlugin, scrypted_sdk.Settings, scrypted_sdk.DeviceProv
 
             for r in objects:
                 obj = Prediction(
-                    r["classId"].astype(float),
-                    r["confidence"].astype(float),
+                    r["classId"],
+                    r["confidence"],
                     Rectangle(
-                        r["xmin"].astype(float),
-                        r["ymin"].astype(float),
-                        r["xmax"].astype(float),
-                        r["ymax"].astype(float),
+                        r["xmin"],
+                        r["ymin"],
+                        r["xmax"],
+                        r["ymax"],
                     ),
                 )
                 objs.append(obj)
@@ -275,9 +275,9 @@ class CoreMLPlugin(PredictPlugin, scrypted_sdk.Settings, scrypted_sdk.DeviceProv
             ),
         )
 
-        coordinatesList = out_dict["coordinates"].astype(float)
+        coordinatesList = out_dict["coordinates"]
 
-        for index, confidenceList in enumerate(out_dict["confidence"].astype(float)):
+        for index, confidenceList in enumerate(out_dict["confidence"]):
             values = confidenceList
             maxConfidenceIndex = max(range(len(values)), key=values.__getitem__)
             maxConfidence = confidenceList[maxConfidenceIndex]

plugins/openvino/src/common/yolo.py

@@ -12,11 +12,11 @@ def parse_yolov10(results, threshold = defaultThreshold, scale = None, confidenc
     for indices in keep:
         class_id = indices[0]
         index = indices[1]
-        confidence = results[class_id + 4, index].astype(float)
-        l = results[0][index].astype(float)
-        t = results[1][index].astype(float)
-        r = results[2][index].astype(float)
-        b = results[3][index].astype(float)
+        confidence = results[class_id + 4, index]
+        l = results[0][index]
+        t = results[1][index]
+        r = results[2][index]
+        b = results[3][index]
         if scale:
             l = scale(l)
             t = scale(t)
@@ -47,7 +47,7 @@ def parse_yolo_nas(predictions):
         pred_cls_label = j[:]
         for box, conf, label in zip(pred_bboxes, pred_cls_conf, pred_cls_label):
             obj = Prediction(
-                int(label), conf.astype(float), Rectangle(box[0].astype(float), box[1].astype(float), box[2].astype(float), box[3].astype(float))
+                int(label), conf, Rectangle(box[0], box[1], box[2], box[3])
             )
             objs.append(obj)
     return objs
@@ -58,11 +58,11 @@ def parse_yolov9(results, threshold = defaultThreshold, scale = None, confidence
     for indices in keep:
         class_id = indices[0]
         index = indices[1]
-        confidence = results[class_id + 4, index].astype(float)
-        x = results[0][index].astype(float)
-        y = results[1][index].astype(float)
-        w = results[2][index].astype(float)
-        h = results[3][index].astype(float)
+        confidence = results[class_id + 4, index]
+        x = results[0][index]
+        y = results[1][index]
+        w = results[2][index]
+        h = results[3][index]
         if scale:
             x = scale(x)
             y = scale(y)
@@ -190,12 +190,12 @@ def parse_yolo_region(blob, original_im_shape, anchors, sigmoid = True):
                 ymax = y + height /2 
                 objects.append(
                     {
-                        'xmin': xmin.astype(float),
-                        'xmax': xmax.astype(float),
-                        'ymin': ymin.astype(float),
-                        'ymax': ymax.astype(float),
-                        'confidence': confidence.astype(float),
-                        'classId': class_id.astype(float),
+                        'xmin': xmin,
+                        'xmax': xmax,
+                        'ymin': ymin,
+                        'ymax': ymax,
+                        'confidence': confidence,
+                        'classId': class_id,
                     }
                 )
 

plugins/openvino/src/ov/__init__.py

@@ -342,7 +342,7 @@ class OpenVINOPlugin(
                 return objs
 
             output = infer_request.get_output_tensor(0)
-            for values in output.data[0][0].astype(float):
+            for values in output.data[0][0]:
                 valid, index, confidence, l, t, r, b = values
                 if valid == -1:
                     break

plugins/openvino/src/predict/__init__.py

@@ -14,12 +14,20 @@ from scrypted_sdk.types import (ObjectDetectionResult, ObjectDetectionSession,
 
 import common.colors
 from detect import DetectPlugin
+from predict.rectangle import Rectangle
 
 class Prediction:
-    def __init__(self, id: int, score: float, bbox: Tuple[float, float, float, float], embedding: str = None):
-        self.id = id
-        self.score = score
-        self.bbox = bbox
+    def __init__(self, id: int, score: float, bbox: Rectangle, embedding: str = None):
+        # these may be numpy values. sanitize them.
+        self.id = int(id)
+        self.score = float(score)
+        # ensure all floats from numpy
+        self.bbox = Rectangle(
+            float(bbox.xmin),
+            float(bbox.ymin),
+            float(bbox.xmax),
+            float(bbox.ymax),
+        )
         self.embedding = embedding
 
 class PredictPlugin(DetectPlugin):