Systems and methods for presenting image classification results

A system for classifying images, comprising:

• one or more processors;
• and
• one or more storage devices storing instructions that, when executed, configure the one or more processors to perform operations comprising:
• executing an augmented reality application, the augmented reality application configured for displaying modified video feeds comprising icons or information superimposed on background images;
• identifying attributes of a captured image using a classification model, the classification model comprising a convolutional neural network, the captured image being a frame of a video feed displayed in the augmented reality application;
• identifying first results based on the identified attributes, the first results being associated with probability scores;
• selecting a subset of the first results based on corresponding probability scores, the first results in the subset having an accumulated probability score greater than a threshold probability score;
• generating a first graphical user interface, the first graphical user interface comprising:
• interactive icons corresponding to first results in the subset,
• at least one input icon, and
• a button;
• and
• upon receiving a selection of the button:
• performing a search to identify second results, the search being based on at least one of selected interactive icons or input in the at least one input icon;
• and
• generating a second graphical user interface, different from the first graphical user interface, displaying the second results;
• determining whether the at least one input icon is empty;
• transmitting, to a server, the captured image and input in the at least one input icon when the at least one input icon is determined not to be empty;
• and
• receiving, from the server, a patch for the classification model, the patch comprising updated model parameters and a classification model exception for the identified attributes.

A computer-implemented method for classifying images, the method comprising:

• executing an augmented reality application, the augmented reality application configured for displaying modified video feeds comprising icons or information superimposed on background images;
• identifying attributes of a captured image using a classification model, the classification model comprising a convolutional neural network, the captured image being a frame of a video feed displayed in the augmented reality application;
• identifying first results based on the identified attributes, the first results being associated with probability scores;
• selecting a subset of the first results based on corresponding probability scores, the first results in the subset having an accumulated probability score greater than a threshold probability score;
• generating a first graphical user interface, the first graphical user interface comprising:
• interactive icons corresponding to first results in the subset,
• at least one input icon, and
• a first button;
• and
• upon receiving a selection of the first button:
• performing a search to identify second results, the search being based on at least one of selected interactive icons or input in the at least one input icon;
• and
• generating a second graphical user interface, different from the first graphical user interface, displaying the second results;
• determining whether the at least one input icon is empty;
• transmitting, to a server, the captured image and input in the at least one input icon when the at least one input icon is determined not to be empty;
• and
• receiving, from the server, a patch for the classification model, the patch comprising updated model parameters and a classification model exception for the identified attributes.