Method and system for vehicle image classification
A method is disclosed that includes operations of receiving user input that includes a vehicle identifier (ID), responsive to receiving the user input, (1) determining a set of vehicle images corresponding to the vehicle ID, (2) a set of classification IDs corresponding to the set of vehicle images, wherein the set of classification IDs includes a classification ID for each vehicle image of the set of vehicle images, and (3) determining feature content corresponding to the set of classification IDs, generating an image-to-feature data map configured to associate the following (i) each vehicle image, (ii) one or more portions of the feature content, and (iii) one or more classification IDs, and transmitting the image-to-feature association to a logic module embedded in webpage code of a webpage, wherein the image-to-feature association includes instructions that, upon execution, cause a rendering of a first vehicle image and first feature content on the webpage.
1. A non-transitory storage medium having logic stored thereon, the logic being executable by one or more processors to perform operations including:
performing analyses of each vehicle image of a collection of vehicle images in accordance with a machine learning model, the analyses of each vehicle image includes (i) detecting a set of points of interest associated with each vehicle image of the collection of vehicle images and (ii) assigning a classification identifier (ID) to each vehicle image of the collection of vehicle images based on the detected set of points of interest, the classification ID for each vehicle image operates as an item type descriptor to indicate that a vehicle image illustrates a particular feature;
receiving user input that includes a vehicle identifier (ID), the vehicle ID being different from the classification ID assigned to each vehicle image;
responsive to receiving the user input, (1) determining a set of vehicle images corresponding to the vehicle ID, (2) determining a set of classification IDs corresponding to the set of vehicle images, wherein each classification ID of the set of classification IDs is associated with a vehicle image of the set of vehicle images based on which features are captured by the vehicle image, and (3) determining feature content corresponding to each classification ID of the set of classification IDs, wherein the feature content includes at least textual information describing the particular feature;
generating an image-to-feature data association for each of the set of vehicle images including a first vehicle image, the image-to-feature data association for the first vehicle image to associate the first vehicle image of the set of vehicle images with (i) one or more portions of the feature content, and (ii) one or more classification IDs representing one or more features captured by the first vehicle image; and
transmitting the image-to-feature association to a logic module embedded in webpage code of a webpage, wherein the image-to-feature association includes instructions that, upon execution, cause a rendering of the first vehicle image and first feature content on the webpage, wherein the first vehicle image and the first feature content are associated with a first classification ID of the set of classification IDs corresponding to the vehicle ID.
2. The non-transitory storage medium of claim 1 , wherein the logic being executable by the one or more processors to perform further operations including:
prior to receiving the user input, performing a machine learning analysis trained on a collection of vehicle images, the collection of vehicle images including the set of vehicle images.
3. The non-transitory storage medium of claim 2 , wherein the machine learning analysis includes applying a machine learning model to the collection of vehicle images, the machine learning model is configured to detect the set of points of interest within each vehicle image and assign the classification ID to each vehicle image based on the detected set of points of interest.
4. The non-transitory storage medium of claim 3 , wherein the collection of vehicle images each having at least one assigned classification ID are stored in a data store.
5. The non-transitory storage medium of claim 3 , wherein the user input includes a dealership ID, and the collection of vehicle images is stored in the data store such that each vehicle image (i) has an assigned classification ID, and (ii) is associated with at least one dealership ID.
6. The non-transitory storage medium of claim 3 , wherein applying the machine learning model includes:
determining one or more features illustrated in the first vehicle image based on a set of detected set of points of interest for the first vehicle image; and
determining the first classification ID based an association of features and classification IDs.
7. The non-transitory storage medium of claim 1 , wherein each classification ID of the set of classification IDs indicates a particular feature illustrated in a vehicle image.
8. The non-transitory storage medium of claim 1 , wherein the vehicle ID is a vehicle identification number (VIN).
9. A method comprising:
performing analyses of each vehicle image of a collection of vehicle images in accordance with a machine learning model, the analyses of each vehicle image includes (i) detecting a set of points of interest associated with each vehicle image of the collection of vehicle images and (ii) assigning a classification identifier (ID) to each vehicle image of the collection of vehicle images based on the detected set of points of interest, a classification ID for each vehicle image operating as an item type descriptor to indicate that a vehicle image illustrates a particular feature;
receiving user input that includes a vehicle identifier (ID), the vehicle ID being different from the classification ID assigned to each vehicle image;
responsive to receiving the user input, (1) determining a set of vehicle images corresponding to the vehicle ID, (2) determining a set of classification IDs corresponding to the set of vehicle images, wherein one or more classification IDs of the set of classification IDs are associated with a first vehicle image of the set of vehicle images based on which features are captured by the first vehicle image, and (3) determining feature content corresponding to each of the set of classification IDs, wherein the feature content includes at least textual information describing the particular feature;
generating an image-to-feature data association for each of the set of vehicle images including a first vehicle image, the image-to-feature data for the first vehicle image to associate the first vehicle image of the set of vehicle images with (i) one or more portions of the feature content, and (ii) the one or more classification IDs representing one or more features captured by the first vehicle image; and
transmitting the image-to-feature association to a logic module embedded in webpage code of a webpage, wherein the image-to-feature association includes instructions that, upon execution, cause a rendering of the first vehicle image and first feature content on the webpage, wherein the first vehicle image and the first feature content are associated with a first classification ID of the set of classification IDs corresponding to the vehicle ID.
10. The method of claim 9 , wherein
the machine learning model is trained on a collection of vehicle images, the collection of vehicle images including the set of vehicle images.
11. The method of claim 10 , wherein the analyses comprise applying the machine learning model to the collection of vehicle images to detect the set of points of interest within each vehicle image and assign the classification ID to each vehicle image based on the detected set of points of interest.
12. The method of claim 11 , wherein the collection of vehicle images each having an assigned classification ID are stored in a data store.
13. The method of claim 11 , wherein the user input includes a dealership ID, and the collection of vehicle images is stored in the data store such that each vehicle image (i) has an assigned classification ID, and (ii) is associated with at least one dealership ID.
14. The method of claim 11 , wherein applying the machine learning model includes:
determining one or more features illustrated in the first vehicle image based on the detected set of points of interest for the first vehicle image.
15. The method of claim 9 , wherein each classification ID of the set of classification IDs indicates a particular feature illustrated in a vehicle image.
16. The method of claim 9 , wherein the vehicle ID is a vehicle identification number (VIN).
17. A system comprising:
one or more processors;
a non-transitory computer-readable medium communicatively coupled to the one or more processors and having logic stored thereon, the logic including instructions being executable by the one or more processors to perform operations including:
performing analyses of each vehicle image of a collection of vehicle images in accordance with a machine learning model, the analyses of each vehicle image includes (i) detecting a set of points of interest associated with each vehicle image of a collection of vehicle images and (ii) assigning a classification identifier (ID) to each vehicle image of the collection of vehicle images based on the detected set of points of interest, a classification ID for each vehicle image operating as an item type descriptor to indicate that a vehicle image illustrates a particular feature;
receiving user input that includes a vehicle identifier (ID), the vehicle ID being different from the classification ID assigned to each vehicle image;
responsive to receiving the user input, (1) determining a set of vehicle images corresponding to the vehicle ID, (2) determining a set of classification IDs corresponding to the set of vehicle images, wherein one or more classification IDs of the set of classification IDs are associated with a first vehicle image of the set of vehicle images based on which features are captured by the first vehicle image, and (3) determining feature content corresponding to the set of classification IDs, wherein the feature content includes at least textual information describing the particular feature;
generating an image-to-feature data association for each of the set of vehicle images including a first vehicle image, the image-to-feature data for the first vehicle image to associate the first vehicle image of the set of vehicle images with (i) one or more portions of the feature content, and (ii) one or more classification IDs representing one or more features captured by the first vehicle image; and
transmitting the image-to-feature association to a logic module embedded in webpage code of a webpage, wherein the image-to-feature association includes instructions that, upon execution, cause a rendering of the first vehicle image and first feature content on the webpage, wherein the first vehicle image and the first feature content are associated with a first classification ID of the set of classification IDs corresponding to the vehicle ID.
18. The system of claim 17 , wherein the
machine learning model being trained on a collection of vehicle images, the collection of vehicle images including the set of vehicle images.
19. The system of claim 17 , wherein the analyses of each vehicle image of the collection of vehicle images in accordance with a machine learning model includes applying the machine learning model to the collection of vehicle images to detect the set of points of interest within each vehicle image and assigns one or more classification IDs to each vehicle image based on the detected set of points of interest.
20. The system of claim 17 , wherein the collection of vehicle images each having an assigned classification ID are stored in a data store.
21. The non-transitory storage medium of claim 1 , wherein each classification ID of the one or more classification IDS further representing the one or more features captured by the first vehicle image further represent a view captured by the first vehicle image.
22. The method of claim 9 , wherein the one or more classification IDs of the set of classification IDS associated with the first vehicle image are further associated with a view captured by the first vehicle image.
23. The method of claim 22 , wherein the feature content further includes textual information describing the view.
24. The system of claim 17 , wherein the one or more classification IDs of the set of classification IDS are associated with the first vehicle image based further on which view is captured by the first vehicle image.
25. The system of claim 24 , wherein the feature content further includes textual information describing the view.