object_detection_base
ObjectDetectionBaseOnnxRoboflowInferenceModel
¶
Bases: OnnxRoboflowInferenceModel
Roboflow ONNX Object detection model. This class implements an object detection specific infer method.
Source code in inference/core/models/object_detection_base.py
28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 |
|
infer(image, class_agnostic_nms=DEFAULT_CLASS_AGNOSTIC_NMS, confidence=DEFAULT_CONFIDENCE, disable_preproc_auto_orient=False, disable_preproc_contrast=False, disable_preproc_grayscale=False, disable_preproc_static_crop=False, iou_threshold=DEFAULT_IOU_THRESH, fix_batch_size=False, max_candidates=DEFAULT_MAX_CANDIDATES, max_detections=DEFAUlT_MAX_DETECTIONS, return_image_dims=False, **kwargs)
¶
Runs object detection inference on one or multiple images and returns the detections.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
image |
Any
|
The input image or a list of images to process. - can be a BGR numpy array, filepath, InferenceRequestImage, PIL Image, byte-string, etc. |
required |
class_agnostic_nms |
bool
|
Whether to use class-agnostic non-maximum suppression. Defaults to False. |
DEFAULT_CLASS_AGNOSTIC_NMS
|
confidence |
float
|
Confidence threshold for predictions. Defaults to 0.5. |
DEFAULT_CONFIDENCE
|
iou_threshold |
float
|
IoU threshold for non-maximum suppression. Defaults to 0.5. |
DEFAULT_IOU_THRESH
|
fix_batch_size |
bool
|
If True, fix the batch size for predictions. Useful when the model requires a fixed batch size. Defaults to False. |
False
|
max_candidates |
int
|
Maximum number of candidate detections. Defaults to 3000. |
DEFAULT_MAX_CANDIDATES
|
max_detections |
int
|
Maximum number of detections after non-maximum suppression. Defaults to 300. |
DEFAUlT_MAX_DETECTIONS
|
return_image_dims |
bool
|
Whether to return the dimensions of the processed images along with the predictions. Defaults to False. |
False
|
disable_preproc_auto_orient |
bool
|
If true, the auto orient preprocessing step is disabled for this call. Default is False. |
False
|
disable_preproc_contrast |
bool
|
If true, the auto contrast preprocessing step is disabled for this call. Default is False. |
False
|
disable_preproc_grayscale |
bool
|
If true, the grayscale preprocessing step is disabled for this call. Default is False. |
False
|
disable_preproc_static_crop |
bool
|
If true, the static crop preprocessing step is disabled for this call. Default is False. |
False
|
*args |
Variable length argument list. |
required | |
**kwargs |
Arbitrary keyword arguments. |
{}
|
Returns:
Type | Description |
---|---|
Any
|
Union[List[ObjectDetectionInferenceResponse], ObjectDetectionInferenceResponse]: One or multiple object detection inference responses based on the number of processed images. Each response contains a list of predictions. If |
Raises:
Type | Description |
---|---|
ValueError
|
If batching is not enabled for the model and more than one image is passed for processing. |
Source code in inference/core/models/object_detection_base.py
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 |
|
make_response(predictions, img_dims, class_filter=None, *args, **kwargs)
¶
Constructs object detection response objects based on predictions.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
predictions |
List[List[float]]
|
The list of predictions. |
required |
img_dims |
List[Tuple[int, int]]
|
Dimensions of the images. |
required |
class_filter |
Optional[List[str]]
|
A list of class names to filter, if provided. |
None
|
Returns:
Type | Description |
---|---|
List[ObjectDetectionInferenceResponse]
|
List[ObjectDetectionInferenceResponse]: A list of response objects containing object detection predictions. |
Source code in inference/core/models/object_detection_base.py
92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 |
|
postprocess(predictions, preproc_return_metadata, class_agnostic_nms=DEFAULT_CLASS_AGNOSTIC_NMS, confidence=DEFAULT_CONFIDENCE, iou_threshold=DEFAULT_IOU_THRESH, max_candidates=DEFAULT_MAX_CANDIDATES, max_detections=DEFAUlT_MAX_DETECTIONS, return_image_dims=False, **kwargs)
¶
Postprocesses the object detection predictions.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
predictions |
ndarray
|
Raw predictions from the model. |
required |
img_dims |
List[Tuple[int, int]]
|
Dimensions of the images. |
required |
class_agnostic_nms |
bool
|
Whether to apply class-agnostic non-max suppression. Default is False. |
DEFAULT_CLASS_AGNOSTIC_NMS
|
confidence |
float
|
Confidence threshold for filtering detections. Default is 0.5. |
DEFAULT_CONFIDENCE
|
iou_threshold |
float
|
IoU threshold for non-max suppression. Default is 0.5. |
DEFAULT_IOU_THRESH
|
max_candidates |
int
|
Maximum number of candidate detections. Default is 3000. |
DEFAULT_MAX_CANDIDATES
|
max_detections |
int
|
Maximum number of final detections. Default is 300. |
DEFAUlT_MAX_DETECTIONS
|
Returns:
Type | Description |
---|---|
List[ObjectDetectionInferenceResponse]
|
List[ObjectDetectionInferenceResponse]: The post-processed predictions. |
Source code in inference/core/models/object_detection_base.py
144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 |
|
predict(img_in, **kwargs)
¶
Runs inference on the ONNX model.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
img_in |
ndarray
|
The preprocessed image(s) to run inference on. |
required |
Returns:
Type | Description |
---|---|
Tuple[ndarray]
|
Tuple[np.ndarray]: The ONNX model predictions. |
Raises:
Type | Description |
---|---|
NotImplementedError
|
This method must be implemented by a subclass. |
Source code in inference/core/models/object_detection_base.py
263 264 265 266 267 268 269 270 271 272 273 274 275 |
|
preprocess(image, disable_preproc_auto_orient=False, disable_preproc_contrast=False, disable_preproc_grayscale=False, disable_preproc_static_crop=False, fix_batch_size=False, **kwargs)
¶
Preprocesses an object detection inference request.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
request |
ObjectDetectionInferenceRequest
|
The request object containing images. |
required |
Returns:
Type | Description |
---|---|
Tuple[ndarray, PreprocessReturnMetadata]
|
Tuple[np.ndarray, List[Tuple[int, int]]]: Preprocessed image inputs and corresponding dimensions. |
Source code in inference/core/models/object_detection_base.py
195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 |
|