def main():
parser = make_parser()
args = parser.parse_args()
# loading the mscoco validation set .
dataType = args.dataType
dataDir = args.dataDir
annFile = '{}/annotations/instances_{}.json'.format(dataDir, dataType)
coco = COCO(annFile)
cats = coco.loadCats(coco.getCatIds())
#nms=[cat['name'] for cat in cats]
imgDir = '{}/images/{}'.format(dataDir, dataType)
#print('COCO categories: \n{}\n'.format(' '.join(nms)))
#nms = set([cat['supercategory'] for cat in cats])
#print('COCO supercategories: \n{}'.format(' '.join(nms)))
coco_demo = COCODemo(cfg, min_image_size=800, confidence_threshold=0.5)
coco_dataset = COCODataset(annFile, dataDir + '/images/' + args.dataType,
True)
initial_detection(coco_dataset,
coco_demo,
args,
max_images=args.max_images,
savedir=args.save_dir)
print(args)
#masked_detection( coco_dataset, coco_demo, args)
masked_detection_2(coco_dataset, coco_demo, args)
print('all_at_once:', args.all_at_once)
def do_coco_evaluation(
dataset,
predictions,
box_only,
output_folder,
iou_types,
expected_results,
expected_results_sigma_tol,
test_result_filename,
):
logger = logging.getLogger("maskrcnn_benchmark.inference")
logger.info("Converting annotations to COCO format...")
coco_annotation_dict = convert_abstract_to_coco(dataset)
dataset_name = dataset.__class__.__name__
coco_annotation_path = os.path.join(output_folder, dataset_name + ".json")
logger.info("Saving annotations to %s" % coco_annotation_path)
with open(coco_annotation_path, "w") as f:
json.dump(coco_annotation_dict, f, indent=2)
logger.info("Loading annotations as COCODataset")
coco_dataset = COCODataset(
ann_file=coco_annotation_path,
root="",
remove_images_without_annotations=False,
transforms=None, # transformations should be already saved to the json
)
return orig_evaluation(
dataset=coco_dataset,
predictions=predictions,
box_only=box_only,
output_folder=output_folder,
iou_types=iou_types,
expected_results=expected_results,
expected_results_sigma_tol=expected_results_sigma_tol,
test_result_filename=test_result_filename,
)
import os
from maskrcnn_benchmark.data.datasets.coco import COCODataset
dataset = COCODataset(
"./my_dataset/train/instances_post_train2018.json",
"./my_dataset/train/images",
True,
)
# boxlist = dataset[0]
# print(boxlist.bbox)
# boxlist = dataset[1][1]
# print(boxlist.bbox)
continue
cutsize = max(x2 - x1, y2 - y1) * 8 / 7 / 2
midx = (x1 + x2) / 2
midy = (y1 + y2) / 2
crop_img = image.crop((int(midx - cutsize), int(midy - cutsize), int(midx + cutsize), int(midy + cutsize)))
closeup.append(crop_img)
return closeup, closeup_target
imgdirs = ['datasets/coco/train2014', 'datasets/coco/val2014']
annofiles = ["datasets/coco/annotations/instances_train2014_base.json", "datasets/coco/annotations/instances_valminusminival2014_base.json"]
if not os.path.exists('datasets/coco/Crops'):
os.mkdir('datasets/coco/Crops')
else:
shutil.rmtree('datasets/coco/Crops')
os.mkdir('datasets/coco/Crops')
for cls in CloseupDataset.CLASSES_COCO_BASE:
os.mkdir('datasets/coco/Crops/' + cls)
cls_count = {cls: 0 for cls in CloseupDataset.CLASSES_COCO_BASE}
for s in range(2):
dataset = COCODataset(annofiles[s], imgdirs[s], True)
for index in range(len(dataset)):
img, annos, _ = dataset.__getitem__(index)
crops, crop_labels = get_closeup(img, annos)
for crop, label in list(zip(crops, crop_labels)):
label = CloseupDataset.CLASSES_COCO_BASE[label]
cls_count[label] += 1
crop.save('datasets/coco/Crops/%s/%d.jpg'%(label, cls_count[label]))
print(cls_count)
print('crop amount:%d'%sum(list(cls_count.values())))