Spaces:
Running
Running
File size: 8,704 Bytes
f965db0 84f01ec f965db0 e599283 84f01ec f965db0 84f01ec f965db0 3359d6e 84f01ec 3359d6e f965db0 84f01ec f965db0 84f01ec f965db0 84f01ec f965db0 84f01ec f965db0 84f01ec f965db0 84f01ec f965db0 84f01ec 3359d6e f965db0 84f01ec 2931c23 84f01ec f965db0 84f01ec |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 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 |
# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""TODO: Add a description here."""
from typing import List, Tuple, Literal
import evaluate
import datasets
import numpy as np
from seametrics.detection import PrecisionRecallF1Support
from seametrics.detection.utils import payload_to_det_metric
from seametrics.payload import Payload
_CITATION = """\
@InProceedings{coco:2020,
title = {Microsoft {COCO:} Common Objects in Context},
authors={Tsung{-}Yi Lin and
Michael Maire and
Serge J. Belongie and
James Hays and
Pietro Perona and
Deva Ramanan and
Piotr Dollar and
C. Lawrence Zitnick},
booktitle = {Computer Vision - {ECCV} 2014 - 13th European Conference, Zurich,
Switzerland, September 6-12, 2014, Proceedings, Part {V}},
series = {Lecture Notes in Computer Science},
volume = {8693},
pages = {740--755},
publisher = {Springer},
year={2014}
}
"""
_DESCRIPTION = """\
This evaluation metric is designed to give provide object detection metrics at different object size levels.
It is based on a modified version of the commonly used COCO-evaluation metrics.
"""
_KWARGS_DESCRIPTION = """
Calculates object detection metrics given predicted and ground truth bounding boxes for a single image.
Args:
predictions: list of predictions for each image. Each prediction should
be a dict containing the following
- 'boxes': list of bounding boxes, xywh in absolute pixel values
- 'labels': list of labels for each bounding box
- 'scores': list of scores for each bounding box
references: list of ground truth annotations for each image. Each reference should
be a dict containing the following
- 'boxes': list of bounding boxes, xywh in absolute pixel values
- 'labels': list of labels for each bounding box
- 'area': list of areas for each bounding box
Returns:
dict containing dicts for each specified area range with following items:
'range': specified area with [max_px_area, max_px_area]
'iouThr': min. IOU-threshold of a prediction with a ground truth box
to be considered a correct prediction
'maxDets': maximum number of detections
'tp': number of true positive (correct) predictions
'fp': number of false positive (incorrect) predictions
'fn': number of false negative (missed) predictions
'duplicates': number of duplicate predictions
'precision': best possible score = 1, worst possible score = 0
large if few false positive predictions
formula: tp/(fp+tp)
'recall' best possible score = 1, worst possible score = 0
large if few missed predictions
formula: tp/(tp+fn)
'f1': best possible score = 1, worst possible score = 0
trades off precision and recall
formula: 2*(precision*recall)/(precision+recall)
'support': number of ground truth bounding boxes considered in the evaluation,
'fpi': number of images with no ground truth but false positive predictions,
'nImgs': number of images considered in evaluation
Examples:
>>> import evaluate
>>> from seametrics.payload import PayloadProcessor
>>> payload = PayloadProcessor(...).payload
>>> for model in payload.models:
>>> module = evaluate.load("SEA-AI/det-metrics", ...)
>>> module.add_from_payload(payload)
>>> result = module.compute()
>>> print(result)
{'all': {
'range': [0, 10000000000.0],
'iouThr': '0.00',
'maxDets': 100,
'tp': 1,
'fp': 3,
'fn': 1,
'duplicates': 0,
'precision': 0.25,
'recall': 0.5,
'f1': 0.3333333333333333,
'support': 2,
'fpi': 0,
'nImgs': 2
}
}
"""
@evaluate.utils.file_utils.add_start_docstrings(_DESCRIPTION, _KWARGS_DESCRIPTION)
class DetectionMetric(evaluate.Metric):
def __init__(
self,
area_ranges_tuples: List[Tuple[str, List[int]]] = [("all", [0, 1e5**2])],
iou_threshold: List[float] = [1e-10],
class_agnostic: bool = True,
bbox_format: str = "xywh",
iou_type: Literal["bbox", "segm"] = "bbox",
**kwargs
):
super().__init__(**kwargs)
area_ranges = [v for _, v in area_ranges_tuples]
area_ranges_labels = [k for k, _ in area_ranges_tuples]
iou_threshold = (
[iou_threshold] if not isinstance(iou_threshold, list) else iou_threshold
)
self.coco_metric = PrecisionRecallF1Support(
iou_thresholds=iou_threshold,
area_ranges=area_ranges,
area_ranges_labels=area_ranges_labels,
class_agnostic=class_agnostic,
iou_type=iou_type,
box_format=bbox_format,
)
def _info(self):
return evaluate.MetricInfo(
# This is the description that will appear on the modules page.
module_type="metric",
description=_DESCRIPTION,
citation=_CITATION,
inputs_description=_KWARGS_DESCRIPTION,
# This defines the format of each prediction and reference
features=datasets.Features(
{
"predictions": [
datasets.Features(
{
"boxes": datasets.Sequence(
datasets.Sequence(datasets.Value("float"))
),
"labels": datasets.Sequence(datasets.Value("int64")),
"scores": datasets.Sequence(datasets.Value("float")),
}
)
],
"references": [
datasets.Features(
{
"boxes": datasets.Sequence(
datasets.Sequence(datasets.Value("float"))
),
"labels": datasets.Sequence(datasets.Value("int64")),
"area": datasets.Sequence(datasets.Value("float")),
}
)
],
}
),
# Additional links to the codebase or references
codebase_urls=[
"https://github.com/SEA-AI/seametrics/tree/main",
"https://lightning.ai/docs/torchmetrics/stable/detection/mean_average_precision.html",
],
)
def add(self, *, prediction, reference, **kwargs):
"""Adds a batch of predictions and references to the metric"""
self.coco_metric.update(prediction, reference)
# does not impact the metric, but is required for the interface x_x
super(evaluate.Metric, self).add(
prediction=[self._np_to_lists(p) for p in prediction],
references=[self._np_to_lists(r) for r in reference],
**kwargs
)
def _np_to_lists(self, d):
"""datasets does not support numpy arrays for type checking"""
for k, v in d.items():
if isinstance(v, dict):
self._np_to_lists(v)
elif isinstance(v, np.ndarray):
d[k] = v.tolist()
return d
def _compute(self, *, predictions, references, **kwargs):
"""Returns the scores"""
return self.coco_metric.compute()["metrics"]
def add_from_payload(self, payload: Payload, model_name: str = None):
"""Converts the payload to the format expected by the metric"""
predictions, references = payload_to_det_metric(payload, model_name)
self.add(prediction=predictions, reference=references)
return self
|