Image classification and labeling

1. A computer-implemented method of classifying images using one or more image classification models, the method comprising:

obtaining training images associated with labels, wherein one or more training images of the training images are associated with two or more labels of the labels, each label corresponding to an image classification class, the labels having a hierarchical structure;

training at least two convolutional neural networks using the training images and the hierarchically-structured labels associated with the training images, a separate convolutional neural network being trained for each level of the hierarchical structure; and

classifying an input image into two or more classes based on the trained at least two convolutional neural networks, the classifying including

applying at least one trained convolutional neural network to a top level of the hierarchical structure, a probability score of a label for each lower level of the hierarchical structure being multiplied by a probability score of a corresponding label at a higher level of the hierarchical structure.

2. The method of claim 1 , wherein a classification layer of the at least two convolutional neural networks is based on soft-sigmoid activation, the soft-sigmoid activation being a combination of a softmax function and a sigmoid function.

3. The method of claim 1 , wherein the training images and the input images include graphically-designed images.

4. The method of claim 1 , wherein the labels are non-mutually exclusive labels.

5. The method of claim 1 , wherein the labels are codes used by a trademark registration organization.

6. The method of claim 1 , wherein the labels are codes used to classify, design patent images or industrial design images.

7. The method of claim 1 , wherein the labels are available as metadata of the training images associated with the labels.

8. The method of claim 1 , wherein the classifying further includes

tagging or labelling the input image with two or more labels corresponding to the two or more classes.

9. The method of claim 1 , further comprising

pre-processing the training images, the training the at least two convolutional neural networks is based on the pre-processed training images and the labels associated with the training images.

10. A computer-implemented method of training an image classification model, the method comprising:

obtaining training images associated with labels, wherein one or more training images of the training images are associated with two or more labels of the labels, each label corresponding to an image classification class, the labels having a hierarchical structure;

classifying training images into one or more classes using, for each level of the hierarchical structure, at least one convolutional neural network, the classifying including

applying the at least one convolutional neural network to a top level of the hierarchical structure, a probability score of a label for each lower level of the hierarchical structure being multiplied by a probability score of a corresponding label at a higher level of the hierarchical structure;

comparing the classification of the training images against labels associated with the training images; and

updating parameters of each respective convolutional neural network based on the comparison of the classification of the training images against the labels associated with the training images.

11. The method of claim 10 , wherein the training images include graphically-designed images.

12. The method of claim 10 , wherein the labels are codes used by a trademark registration organization.

13. The method of claim 10 , further comprising

pre-processing the training images prior to the classifying.

14. The method of claim 10 , wherein a classification layer of each respective convolutional neural network is based on soft-sigmoid activation, the soft-sigmoid activation being a combination of a softmax function and a sigmoid function.

15. A system for classifying images using one or more image classification models, the system comprising:

a training image database comprising training images associated with labels, one or more training images of the training images being associated with two or more labels of the labels, the two or more labels of the labels having a hierarchical structure; and

processing circuitry communicably coupled to the training image database and configured to

obtain the training images from the training image database, and

train at least two convolutional neural networks using the training images and the hierarchically-structured labels associated with the training images, a separate convolutional neural network being trained for each level of the hierarchical structure, wherein

the at least two convolutional neural networks are configured to classify an input image from an input image database into two or more classes by

applying one or more convolutional neural network of the at least two convolution neural networks to a top level of the hierarchical structure, a probability score of a label for each lower level of the hierarchical structure being multiplied by a probability score of a corresponding label at a higher level of the hierarchical structure.

16. The system of claim 15 , wherein a classification layer of each of the at least two convolutional neural network is based on soft-sigmoid activation, the soft-sigmoid activation being a combination of a softmax function and a sigmoid function.

17. The system of claim 15 , wherein the labels are codes used by a trademark registration organization or an industrial design registration organization.