System and method for toy recognition

A system according to any one of the preceding claims;

• wherein the recognition system is further configured to obtain, based on the predicted object identifier, further object information about the depicted real-world toy object.

A system according to any one of the preceding claims;

• wherein the model creation system is further configured to process at least a first digital image so as to create one or more processed versions of the first digital image;
• and wherein training the convolutional classification model is based on one or more processed versions of the first digital image, each processed version being annotated with an object identifier associated with the first digital image.

The system according to any one of the preceding claims, wherein the convolutional classification model is a convolutional neural network.

The system according to any one of the preceding claims, wherein the convolutional classification model is a deep convolutional classification model.

The system according to any one of the preceding claims wherein the model creation system comprises one or more graphics processing units (“GPU”) configured to execute a training process for training the convolutional classification model.

The system according to any one of the preceding claims, wherein the trained convolutional classification model is configured to output, when presented with a digital image, a prediction indicative of respective likelihoods that said digital image depicts respective real-world toy objects included in the said training database.

The system according to any one of the preceding claims, wherein the recognition system is implemented by a mobile device comprising a digital camera, a display and a processor.

The system according to any one of the preceding claims;

• wherein the model creation system comprises an image capturing device and a support member for receiving a real-world toy object;
• wherein at least one of the support member and the image capturing device is movably arranged so as to vary a viewpoint of the image capturing device relative to the support member;
• and wherein the model creation system is configured to capture multiple images from respective viewpoints of a real-world toy object positioned on the support member.

The system according to any one of the preceding claims;

• wherein the recognition system is further configured to estimate one or more additional attributes of the real-world object depicted in the captured image, in addition to an identification of the real-world object.

The system according to any one of the preceding claims;

• wherein the convolutional classification model comprises at least an upstream and a downstream convolutional layer of feature-extracting kernels, the downstream layer receiving an input based on an output of the upstream layer;
• and wherein the downstream layer has a number of kernels larger than the number of kernels of the upstream layer.

The system according to any one of the preceding claims;

• wherein the convolutional classification model comprises at least 6 convolutional layers of feature-extracting kernels.

The system according to any one of the preceding claims;

• wherein the convolutional classification model comprises a plurality of convolutional layers and one or more pooling layers arranged as a sequence of layers where information is past from an input layer in a downstream direction to a final layer of said sequence;
• wherein the final layer is a 1-dimensional layer.

A computer implemented method for recognizing real-world toy objects from one or more captured images, the method comprising:

• receiving a captured image of a real-world toy object;
• using a trained convolutional classification model to predict a matching object identifier.

A game system according to any one of claims 32 through 37;

• comprising a recognition system having one or more of the features defined in any one of claim 1-26 or 40-41.