partialconv

Partial Convolution

input

Ailia input shape: (1, 3, 224, 224)

output

class_count=5
+ idx=0
  category=731[plunger, plumber's helper ]
  prob=12.340980529785156
+ idx=1
  category=543[dumbbell ]
  prob=11.191944122314453
+ idx=2
  category=680[nipple ]
  prob=10.75782299041748
+ idx=3
  category=422[barbell ]
  prob=10.286931991577148
+ idx=4
  category=844[switch, electric switch, electrical switch ]
  prob=9.976827621459961

usage

Automatically downloads the onnx and prototxt files on the first run. It is necessary to be connected to the Internet while downloading.

For the sample image,

$ python3 partialconv.py

If you want to specify the input image, put the image path after the --input option.

$ python3 partialconv.py --input IMAGE_PATH

By adding the --video option, you can input the video.
If you pass 0 as an argument to VIDEO_PATH, you can use the webcam input instead of the video file.

$ python3 partialconv.py --video VIDEO_PATH

You can select a model from resnet50 | vgg16_bn | pdresnet50 | pdresnet101 | pdresnet152 by adding --arch. Please note that pdresnet152 does not currently executable on Mac OS.

Reference

Partial Convolution Layer for Padding and Image Inpainting

Framework

PyTorch 1.2.0

Model Format

ONNX opset = 10

Netron

resnet50.onnx.prototxt

vgg16_bn.onnx.prototxt

pdresnet50.onnx.prototxt