Data mining using group classifiers

A system comprising:

• one or more processing units to:
• receive training data representative of one or more first images and one or more group classifications associated with the one or more first images, an individual group classification of the one or more group classifications indicating a first group or a second group;
• update, during a first training iteration, one or more parameters associated with a classifier using the training data;
• based at least on the first training iteration being complete, determine, using the classifier and based at least on image data representative of at least a second image, that the second image is associated with the second group;
• and
• generating, based at least on the second image being associated with the second group and by at least adding the image data to the training data, updated training data for training the classifier during a second training iteration.

One or more processors comprising processing circuitry to:

• update, based at least on training image data representative of one or more training images and one or more group classifications associated with the one or more training images, one or more parameters associated with a first classifier, the one or more group classifications including at least a first group classification that includes a plurality of object classifications;
• generate, by the first classifier and based at least on processing image data representative of images, first output data indicating that a first portion of the images is associated with the first group classification and a second portion of the images is associated with a second group classification;
• and
• generate, by a second classifier and based at least on processing a portion of the image data that represents the first portion of the images, an object classification of the plurality of object classification that is associated with at least an image of the first portion of the images.