Data preparation for artificial intelligence models

A method comprising:

• receiving data characterizing a first plurality of images;
• annotating a first subset of images of the first plurality of images based at least in part on a first user input to generate annotated first subset of images, wherein the annotating includes labelling one or more features of the first subset of images;
• generating, by a training code, an annotation code, the training code configured to receive the annotated first subset of images as input and output the annotation code;
• and
• iteratively annotating at least a portion of the received first plurality of images, wherein each iteration of the iterative annotation includes:
• selecting a subset of images from the first plurality of images,
• annotating, by the annotation code, the selected subset of images;
• selecting one or more images from the annotated images;
• providing the selected one or more images to a user;
• receiving a second user input indicative of annotation information associated with the selected one or more images;
• and
• calculating a characteristic metric associated with the selected one or more images based on the second user input.

A method comprising:

• receiving data characterizing a plurality of images;
• calculating a quality metric associated with the plurality of images based on a discrepancy between an image variation characteristic of the plurality of images and predetermined training data requirements of a target deep learning code;
• determining, based on the quality metric, that the plurality of images require image modification;
• and
• generating one or more of a plurality of synthesized images and a plurality of augmented images from the plurality of images based on the quality metric.

A system comprising:

• at least one data processor;
• memory coupled to the at least one data processor, the memory storing instructions to cause the at least one data processor to perform operations comprising:
• receiving data characterizing a first plurality of images;
• annotating a first subset of images of the first plurality of images based at least in part on a first user input to generate annotated first subset of images, wherein the annotating includes labelling one or more features of the first subset of images;
• generating, by a training code, an annotation code, the training code configured to receive the annotated first subset of images as input and output the annotation code, wherein the training and the annotation code includes computer executable instructions;
• and
• annotating, by the annotation code, a second subset of images of the first plurality of images to generate annotated second subset of images, wherein the annotating includes labelling one or more features of the second subset of images.

A computer program product comprising a non-transitory machine-readable medium storing instructions that, when executed by at least one programmable processor that comprises at least one physical core and a plurality of logical cores, cause the at least one programmable processor to perform operations comprising:

• receiving data characterizing a first plurality of images;
• annotating a first subset of images of the first plurality of images based at least in part on a first user input to generate annotated first subset of images, wherein the annotating includes labelling one or more features of the first subset of images;
• generating, by a training code, an annotation code, the training code configured to receive the annotated first subset of images as input and output the annotation code, wherein the training and the annotation code includes computer executable instructions;
• and
• annotating, by the annotation code, a second subset of images of the first plurality of images to generate annotated second subset of images, wherein the annotating includes labelling one or more features of the second subset of images.