how to do when image has no object labeled?

[h030162]

hi! thanks for your great work. target = np.array(target) bbox = target[:, :4] labels = target[:, 4]

when nothing to labeled, target shape is (0,) in this time how to do? thank you again

[AsatledFish]

hi! can you solved this problem?

[h030162]

    bbox=[]
    labels=[]
    emptybbox=[[10,10,100,100]]
    emptylabels=[[0]]
    if len(target) > 0:
        bbox = target[:, :4]
        for i in range(len(bbox)):
            bbox[i, 0] = bbox[i, 0] if bbox[i, 0] < width else width
            bbox[i, 1] = bbox[i, 1] if bbox[i, 1] < height else height
            bbox[i, 2] = bbox[i, 2] if bbox[i, 2] < width else width
            bbox[i, 3] = bbox[i, 3] if bbox[i, 3] < height else height
            for j in range(len(bbox[i])):
                bbox[i,j] = bbox[i,j] if bbox[i,j] > 0 else 0
        labels = target[:, 4]
    if self.transform is not None:
        if len(bbox) > 0:
            #print("len(bbox)", bbox, img.shape)
            annotation = {'image': img, 'bboxes': bbox, 'category_id': labels}
            augmentation = self.transform(**annotation)
            img = augmentation['image']
            bbox = augmentation['bboxes']
            labels = augmentation['category_id']
        else:
            #print("len(bbox)2", bbox)
            annotation1 = {'image': img, 'bboxes': emptybbox, 'category_id': emptylabels}
            augmentation = self.transform(**annotation1)
            img = augmentation['image']
    return {'image': img, 'bboxes': bbox, 'category_id': labels}

it's my code. I'm not sure this is right or not.