System and method for automatically curating and displaying images
The present disclosure is directed to automatically curating a set of images according to a predetermined set of compliance rules and displaying those images. The machine learning and artificial intelligence technology can differentiate between images and identify features within images. The features may include a desired perspective, inconsistent backgrounds, low detail, blurriness, shadows, glare, reflections, unwanted information, unwanted elements, poles, trees, lack or focus, poor resolution, rain, snow, fumes, smoke, mud, unwanted banners, unwanted overlays, etc. The technology is trained to identify these features in the images and automatically tag the images with information relating to the features. A user selects a set of predetermined rules to be applied to the images. The technology then uses the tags to apply these selected rules to the images to modify the images and display the modified images in a predetermined sequence and arrangement.
1. At least one non-transitory computer-readable storage medium having computer-executable instructions stored thereon which, when executed by a computer, causes the computer to perform a method comprising:
providing a first predetermined set of visual display rules associating a first predetermined set of actions with a first type vehicle, the first predetermined set of rules comprising:
a first rule of visually-perceptible image display associated with a first visually-perceptible feature; and
second rule of visually-perceptible image display associated with a second visually-perceptible feature;
providing a second predetermined set of visual display rules associating a second predetermined set of actions with a second type vehicle, the second predetermined set of rules comprising:
a third rule of visually-perceptible image display associated with a third visually˜perceptible feature; and
a fourth rule of visually-perceptible image display associated with a fourth visually-perceptible feature;
providing a first master set of rules comprising rules from the first predetermined set of rules and rules from the second predetermined set of rules;
capturing a first image of a first vehicle of the first type vehicle from a first perspective wherein the first image is not in compliance with the first predetermined set of visual display rules;
capturing a second image of the first vehicle from a second perspective wherein the second image is not in compliance with the first predetermined set of visual display rules;
capturing a third image of a second vehicle of the second type vehicle from a third perspective wherein the third image is not in compliance with the second predetermined set of visual display rules;
capturing a fourth image of the second vehicle from a fourth perspective wherein the fourth image is not in compliance with the second predetermined set of visual display rules;
automatically using machine learning and contextual image classification to identify a first portion of contextual information associated with a first target pixel based on patterns in the first image using groupings of pixels surrounding the first target pixel in the first image;
automatically using machine learning to compare the first portion of contextual information against the master set of rules to identify a first visually-perceptible feature in the first image associated with the master set of rules, generating a first metadata tag based on the first visually-perceptible feature, and associating the first metadata tag with the first image;
automatically using the first metadata tag to execute the first rule and add a first visually-perceptible display element to the first image;
automatically using machine learning and contextual image classification to identify a second portion of contextual information associated with a second target pixel based on patterns in the second image using groupings of pixels surrounding the second target pixel in the second image;
automatically using machine learning to compare the second portion of contextual information against the master set of rules to identify a second visually-perceptible feature in the second image associated the master set of rules, generating a second metadata tag based on the second visually-perceptible feature, and associating the second metadata tag with the second image;
automatically using the second metadata tag to execute the second rule and add a second visually-perceptible display element to the second image;
automatically using machine learning and contextual image classification to identify a third portion of contextual information associated with a third target pixel based on patterns in the third image using groupings of pixels surrounding the third target pixel in the third image;
automatically using machine learning to compare the third portion of contextual information against the master set of rules to identify a third visually-perceptible feature in the third image associated the master set of rules, generating a third metadata tag based on the third visually-perceptible feature, and associating the third metadata tag with the third image;
automatically using the third metadata tag to execute the third rule and add a third visually-perceptible display element to the third image;
automatically using machine learning and contextual image classification to identify a fourth portion of contextual information associated with a fourth target pixel based on patterns in the fourth image using groupings of pixels surrounding the fourth target pixel in the fourth image;
automatically using machine learning to compare the fourth portion of contextual information against the master set of rules to identify a fourth visually-perceptible feature in the fourth image associated the master set of rules, generating a fourth metadata tag based on the fourth visually-perceptible feature, and associating the fourth metadata tag with the fourth image;
automatically using the fourth metadata tag to execute the fourth rule and add a fourth visually-perceptible display element to the fourth image;
automatically displaying the first image with the first visually-perceptible display element in compliance with the first predetermined set of visual display rules;
automatically displaying the second image with the second visually-perceptible display element in compliance with the first predetermined set of visual display rules;
automatically displaying the third image with the third visually-perceptible display element in compliance with the second predetermined set of visual display rules; and
automatically displaying the fourth image with the fourth visually-perceptible display element in compliance with the second predetermined set of visual display rules.
2. The at least one non-transitory computer-readable storage medium of claim 1 , wherein the first type vehicle is a vehicle manufactured by a first manufacturer and wherein the second type vehicle is a vehicle manufactured by a second vehicle manufacturer.
3. The at least one non-transitory computer-readable storage medium of claim 2 , wherein the first predetermined set of rules is a first predetermined set of rules received from the first vehicle manufacturer and wherein the second predetermined set of rules is a second predetermined set of rules received from the second vehicle manufacturer.
4. The at least one non-transitory computer-readable storage medium of claim 1 , wherein automatically identifying a vehicle type of the first image as the first type comprises identifying a plurality of features of the first vehicle in the first image and using the plurality of features to identify the vehicle type of the first image as the first type using machine learning.
5. The at least one non-transitory computer-readable storage medium of claim 1 , wherein the first predetermined set of rules is provided by a first manufacturer of the first vehicle type and wherein the second predetermined set of rules is provided by a second manufacturer of the second vehicle type.
6. The at least one non-transitory computer-readable storage medium of claim 5 , wherein the first predetermined set of rules are first marketing display compliance standards of the first manufacturer and wherein the second predetermined set of rules are second marketing display compliance standards of the second manufacturer.
7. The at least one non-transitory computer-readable storage medium of claim 1 , further comprising automatically using the third visually-perceptible feature to prevent application of any of the second predetermined set of rules to the third image that are not also in the first predetermined set of rules.
8. The at least one non-transitory computer-readable storage medium of claim 1 , wherein the first visually-perceptible feature is a damage location on a vehicle.
9. The at least one non-transitory computer-readable storage medium of claim 8 , further comprising automatically adding an interactive hotspot to the image at the damage location.
10. The at least one non-transitory computer-readable storage medium of claim 1 , wherein the first visually-perceptible feature is a feature selected from the group consisting of a vehicle make, a vehicle model, a vehicle angle, a vehicle year, a vehicle trim, an image perspective, a banner, text, a cropping, a background segmentation, a level of image detail, a shadow, a glare, a reflection, a license plate number, a person, an animal, a tree, a lack of focus, rain, fumes, snow, dust, smoke, mud, a banner, and an overlay.
11. The at least one non-transitory computer-readable storage medium of claim 10 , further comprising automatically adding an interactive hotspot to the first image at the feature location.
12. The at least one non-transitory computer-readable storage medium of claim 11 , further comprising automatically using machine learning and contextual image classification to identify a fifth portion of contextual information associated with the first image using the first visually-perceptible feature.
13. The at least one non-transitory computer-readable storage medium of claim 11 , further comprising automatically using machine learning and contextual image classification to identify a fifth portion of contextual information associated with the first image using automatic segmentation of the image.