Faster_Mask_RCNN

pytorch obeject detection finetuning tutorial

python : 3.7.5

torch : 1.8.1+cu111

pycocotools : 2.0.2

• faster & mask R-CNN 튜닝

1. Struct

Faster_Mask_RCNN
| detection
└─| coco_eval.py
  | coco_utils.py
  │ engine.py
  │ group_by_aspect_ratio.py
  │ presets.py
  │ train.py
  │ transforms.py
  │ utils.py
| PenFudanPed
| active.py
| datasets.py
| networks.py
| run.py

• detection : pytorch에서 기본으로 재공해주는 라이브러리 예시
• PenFudanPed : 학습 및 테스트를 위한 영상 데이터 폴더 : PenFudanPed - PASCAL Annotation Version 1.00
• active.py : Trian/Predict/View
• datasets.py : 데이터 전처리
• networks.py : 모델 생성
• run.py : 실행

2. Setting

• 가상환경(venv) 파일 수정

  • venv\lib\site-packages\torchvision\models\detection\faster_rcnn.py
  • venv\lib\site-packages\torchvision\models\detection\mask_rcnn.py

• import 수정

faster_rcnn.py 수정

...
# 17 line
__all__ = [
    "FasterRCNN", "fasterrcnn_resnet_fpn", "fasterrcnn_resnet50_fpn", "fasterrcnn_mobilenet_v3_large_320_fpn",
    "fasterrcnn_mobilenet_v3_large_fpn"
]

...
# write
def fasterrcnn_resnet_fpn(net='resnet50', pretrained=False, progress=True,
                          num_classes=91, pretrained_backbone=True, trainable_backbone_layers=None, **kwargs):

    trainable_backbone_layers = _validate_trainable_layers(
        pretrained or pretrained_backbone, trainable_backbone_layers, 5, 3)

    backbone = resnet_fpn_backbone(
        net, pretrained_backbone, trainable_layers=trainable_backbone_layers)
    model = FasterRCNN(backbone, num_classes, **kwargs)

    return model

mask_rcnn.py 수정

...
# 13 line
__all__ = [
    "MaskRCNN", "maskrcnn_resnet_fpn", "maskrcnn_resnet50_fpn",
]

...


def maskrcnn_resnet_fpn(net='resnet50', pretrained=False, progress=True,
                        num_classes=91, pretrained_backbone=True, trainable_backbone_layers=None, **kwargs):
    trainable_backbone_layers = _validate_trainable_layers(
        pretrained or pretrained_backbone, trainable_backbone_layers, 5, 3)

    backbone = resnet_fpn_backbone(
        net, pretrained_backbone, trainable_layers=trainable_backbone_layers)
    model = MaskRCNN(backbone, num_classes, **kwargs)

    return model

detection의 import수정( coco_eval.py, coco_utils.py, engine.py, presets.py, train.py )

  import utils => from . import utils
  import transforms as T => from . import transforms as T
  import presets = > from . import presets
  from coco_utils => from .coco_utils
  from coco_eval => from .coco_eval
  from group_by_aspect_ratio => from .group_by_aspect_ratio
  from engine => from .engine

3. git repository

4. Run : python run.py

5. Result

• image
  

• masks