Skip to content

merge

Crops Merge App.

API REFERENCE

crops_merge(ground_truth_file, output_file, predictions_file=None, apply_nms=True, score_thr=0.0, iou_thr=0.5)

Merge and translate ground_truth_file or predictions to ground_truth's old_images coordinates.

Parameters:

Name Type Description Default
ground_truth_file str

Path to COCO ground truth file of crops. Generated with crops_split.

 required
output_file str

Path where the merged annotations will be saved.

 required
predictions_file Optional[str]

Path to COCO predictions file over ground_truth_file. If not None, the annotations of predictions_file will be merged instead of ground_truth_file's.

 None
apply_nms bool

Whether to apply Non Maximum Supression to the merged predictions of each image. Defaults to True.

 True
score_thr float

Predictions bellow score_thr will be filtered. Only used if apply_nms. Defaults to 0.0.

 0.0
iou_thr float

None of the filtered predictions will have an iou above iou_thr to any other. Only used if apply_nms. Defaults to 0.5.

 0.5
Source code in pyodi/apps/crops/crops_merge.py 
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
@logger.catch(reraise=True)
def crops_merge(
    ground_truth_file: str,
    output_file: str,
    predictions_file: Optional[str] = None,
    apply_nms: bool = True,
    score_thr: float = 0.0,
    iou_thr: float = 0.5,
) -> str:
    """Merge and translate `ground_truth_file` or `predictions` to `ground_truth`'s `old_images` coordinates.

    Args:
        ground_truth_file: Path to COCO ground truth file of crops. Generated with
            `crops_split`.
        output_file: Path where the merged annotations will be saved.
        predictions_file: Path to COCO predictions file over `ground_truth_file`.
            If not None, the annotations of predictions_file will be merged instead of ground_truth_file's.
        apply_nms: Whether to apply Non Maximum Supression to the merged predictions of
            each image. Defaults to True.
        score_thr: Predictions bellow `score_thr` will be filtered. Only used if
            `apply_nms`. Defaults to 0.0.
        iou_thr: None of the filtered predictions will have an iou above `iou_thr` to
            any other. Only used if `apply_nms`. Defaults to 0.5.

    """
    ground_truth = json.load(open(ground_truth_file))

    crop_id_to_filename = {x["id"]: x["file_name"] for x in ground_truth["images"]}

    stem_to_original_id = {
        Path(x["file_name"]).stem: x["id"] for x in ground_truth["old_images"]
    }
    stem_to_original_shape = {
        Path(x["file_name"]).stem: (x["width"], x["height"])
        for x in ground_truth["old_images"]
    }

    if predictions_file is not None:
        annotations = json.load(open(predictions_file))
    else:
        annotations = ground_truth["annotations"]

    for n, crop in enumerate(annotations):
        if not n % 10:
            logger.info(n)
        filename = crop_id_to_filename[crop["image_id"]]
        parts = Path(filename).stem.split("_")

        stem = "_".join(parts[:-2])
        original_id = stem_to_original_id[stem]
        crop["image_id"] = original_id

        # Corners are encoded in crop's filename
        # See crops_split.py
        crop_row = int(parts[-1])
        crop_col = int(parts[-2])
        crop["bbox"][0] += crop_col
        crop["bbox"][1] += crop_row
        crop["original_image_shape"] = stem_to_original_shape[stem]

    if apply_nms:
        annotations = nms_predictions(annotations, score_thr=score_thr, iou_thr=iou_thr)
        output_file = str(
            Path(output_file).parent
            / f"{Path(output_file).stem}_{score_thr}_{iou_thr}.json"
        )

    if predictions_file is not None:
        new_ground_truth = annotations
    else:
        new_ground_truth = {
            "info": ground_truth.get("info", []),
            "licenses": ground_truth.get("licenses", []),
            "categories": ground_truth.get("categories", []),
            "images": ground_truth.get("old_images"),
            "annotations": annotations,
        }

    with open(output_file, "w") as f:
        json.dump(new_ground_truth, f, indent=2)

    return output_file