Image Classification

Query image and prediction

BiWAKO.ResNet

Basic ResNet V2 trained on ImageNet

Attributes:

Name	Type	Description
model_path	str	Path to the model file. If automatic download is not enabled, this path is used to save the file.
model	onnxruntime.InferenceSession	Inference session for the model.
input_name	str	Name of the input node.
output_name	str	Name of the output node.
input_shape	tuple	Shape of the input node.
label	dict	Dictionary of the label. The key is the class index and the value is the class name.
mean	np.ndarray	Mean of the normalization.
var	np.ndarray	Variance of the normalization.

__init__(self, model='resnet18v2') special

Initialize ResNet

Available models: "resnet152v2" "resnet101v2" "resnet50v2" or "resnet18v2"

Parameters:

Name	Type	Description	Default
model	str	Choice of the model from the table above or path to the downloaded onnx file. If the file has not been downloaded, the automatic download is triggered. Defaults to "resnet18v2".	'resnet18v2'

predict(self, image)

Return the prediction of the model

Parameters:

Name	Type	Description	Default
image	Image	Image to be predicted. Accept path or cv2 image.	required

Returns:

Type	Description
np.ndarray	1 by 1000 array of the prediction. Softmax is not applied.

render(self, prediction, image, topk=5, **kwargs)

Return the original image with the predicted class names

Parameters:

Name	Type	Description	Default
prediction	np.ndarray	Prediction returned by predict().	required
image	Image	Image to be predicted. Accept path or cv2 image.	required
topk	int	Number of classes to display with higher probability. Defaults to 5.	5

Returns:

Type	Description
np.ndarray	Image with the predicted class names.