Strange behavior of weighted_boxes_fusion

[shonenkov]

@ZFTurbo

boxes = [
    [[410, 464, 354, 410],
    [511, 89, 470, 32],
    [503, 213, 447, 166],
    [300, 444, 252, 391],
    [290, 184, 234, 133]],
    
    [[354, 412, 409, 463],
    [447, 165, 504, 212],
    [251, 392, 299, 444],
    [470, 35, 511, 90],
    [187, 316, 240, 378]],

    [[447, 166, 503, 213],
    [355, 412, 411, 464],
    [470, 33, 511, 88],
    [251, 391, 300, 444],
    [234, 132, 289, 184]],

    [[251, 391, 300, 445],
    [354, 412, 410, 463],
    [447, 166, 503, 212],
    [235, 134, 289, 184],
    [191, 316, 239, 380]],

    [[410, 465, 355, 412],
    [511, 88, 470, 33],
    [504, 213, 448, 167],
    [299, 444, 251, 392],
    [289, 185, 236, 133]]
]

scores = [
    [0.893,0.886,0.865,0.864,0.801],
    [0.915,0.881,0.873,0.852,0.844],
    [0.896, 0.895,0.860,0.855,0.801],
    [0.900,0.889,0.861,0.838,0.831],
    [0.897,0.877, 0.877,0.861,0.820],
]

labels = [
    [1, 1, 1, 1, 1],
    [1, 1, 1, 1, 1],
    [1, 1, 1, 1, 1],
    [1, 1, 1, 1, 1],
    [1, 1, 1, 1, 1]
]

wbf_boxes, wbf_scores, wbf_labels = ensemble_boxes.ensemble_boxes_wbf.weighted_boxes_fusion(boxes, scores, labels, weights=None, iou_thr=0.43, skip_box_thr=0.44)

wbf_boxes.shape
>>> (16, 4)

I think we see excess wbf_boxes, it is very strange. Picture with demonstration:

Python 3.7.6
numpy==1.18.1
ensemble_boxes==1.0.1

[ZFTurbo]

Coordinates for boxes expected to be normalized e.g in range [0; 1]. Order: x1, y1, x2, y2.

I will add warning on such cases.

[shonenkov]

@ZFTurbo no, it doesnt help!

wbf_boxes, wbf_scores, wbf_labels = ensemble_boxes.ensemble_boxes_wbf.weighted_boxes_fusion(
    (np.array(boxes)/511).tolist(),
    scores, 
    labels, 
    weights=None, 
    iou_thr=0.43, 
    skip_box_thr=0.44
)
wbf_boxes.shape
>>> (16, 4)

[ZFTurbo]

x1 must be lower than x2 and y1 must be lower than y2. I will fix it next release.

[shonenkov]

@ZFTurbo thank you! I will try it. I am looking forward to the new release :)

[shonenkov]

@ZFTurbo thank you! it works for me, I have used temporary "hardcode" (maybe it helps anyone):

   def prepare_boxes(boxes):
        result_boxes = boxes.copy()
        result_boxes[:,0] = np.min(boxes[:, [0,2]], axis=1)
        result_boxes[:,2] = np.max(boxes[:, [0,2]], axis=1)
        result_boxes[:,1] = np.min(boxes[:, [1,3]], axis=1)
        result_boxes[:,3] = np.max(boxes[:, [1,3]], axis=1)
        return result_boxes