Vehicle listing image detection and alert system
An image error identification system retrieves an image associated with a vehicle listing and uses various machine learning models to classify the image and generate identification data that may include a vehicle make, model, trim level, and/or various features of the vehicle present in the image. The identification data is compared to the rest of the vehicle listing to detect a mismatch between the image and the vehicle listing. An alert is generated, when a mismatch is detected, indicating the one of the image or the data in the vehicle listing is incorrect.
1. A computer-implemented method comprising:
retrieving an image from a data set corresponding to an individual identifier;
classifying, using a first machine learning model, the image to determine a first portion of identification data for the image;
classifying, using a second machine learning model, the image to determine a second portion of the identification data, in response to successfully classifying the image using the first machine learning model;
comparing at least one portion of the identification data to a member of the data set to determine whether there is a mismatch between the at least one portion of the identification data and the member of the data set; and
generating a mismatch alert to indicate an error in the data set, in response to the mismatch between the at least one portion of the identification data and the member of the data set.
2. The method of claim 1 , wherein the retrieving of the image from the data set further comprises:
retrieving an image identifier from the data set, wherein the image identifier includes a universal resource locator (URL) identifying a local or remote storage location that stores the image; and
retrieving the image corresponding to the image identifier from the local or remote storage location.
3. The method of claim 1 , wherein
the first portion of the identification data comprises a first general image class, and
the first general image class comprises an exterior image class associated with an exterior image of a vehicle.
4. The method of claim 3 , wherein the second portion of the identification data comprises a general identifier for the image,
the general identifier corresponds to one or more sub-classes of a plurality of sub-classes within the first general image class,
the generating of the mismatch alert is in response to a mismatch between the second portion of the identification data comprising the general identifier and the member of the data set, and
the member of the data set comprises an identifier from the plurality of sub-classes.
5. The method of claim 3 , further comprising:
classifying, using a third machine learning model, the image to determine a third portion of the identification data, wherein
the second portion of the identification data comprises a general identifier for the image, and
the third portion of the identification data comprises an indication of a feature in the image.
6. The method of claim 5 , wherein
the individual identifier identifies a vehicle listing and the data set is a vehicle listing data set,
the member of the data set corresponds to the feature in the image, and
the generating of the mismatch alert is in response to a mismatch between the third portion comprising the indication of the feature and the member of the data set corresponding to the feature.
7. The method of claim 1 , further comprising modifying the image from the data set, in response to the mismatch between the at least one portion of the identification data and the member of the data set.
8. The method of claim 1 , further comprising modifying the member of the data set in response to the mismatch between the at least one portion of the identification data and the member of the data set.
9. The method of claim 1 , further comprising:
identifying a first modification, among two or more candidate modifications, the first modification comprising modifying the member of the data set;
identifying a second modification, among the two or more candidate modifications, the second modification comprising replacing the image to change the identification data;
for each of the first modification and the second modification, determining a number of steps required to eliminate mismatches between the identification data and all members of the data set; and
performing a modification, among the two or more candidate modifications, having least number of steps.
10. A system comprising:
a processor; and
a memory coupled to the processor, wherein the processor and the memory are configured to:
retrieve an image from a data set corresponding to an individual identifier;
classify, using a first machine learning model, the image to determine a first portion of identification data for the image;
classify, using a second machine learning model, the image to determine a second portion of the identification data, in response to successfully classifying the image using the first machine learning model;
compare at least one portion of the identification data to a member of the data set to determine whether there is a mismatch between the at least one portion of the identification data and the member of the data set; and
generate a mismatch alert to indicate an error in the data set, in response to the mismatch between the at least one portion of the identification data and the member of the data set.
11. The system of claim 10 , wherein the processor and the memory are further configured to:
retrieve an image identifier from the data set, wherein the image identifier includes a universal resource locator (URL) identifying a local or remote storage location that stores the image; and
retrieve the image corresponding to the image identifier, from the local or remote storage location.
12. The system of claim 10 , wherein
the first portion of the identification data comprises a first general image class, and
the first general image class comprises an exterior image class associated with an exterior image of a vehicle.
13. The system of claim 12 , wherein
the second portion of the identification data comprises a general identifier for the image,
the general identifier corresponds to one or more of a plurality of sub-classes within the first general image class,
the generation of the mismatch alert is in response to a mismatch between the second portion of the identification data comprising the general identifier and the member of the data set, and
the member of the data set comprises an identifier from the plurality of sub-classes.
14. The system of claim 12 , wherein the processor and the memory are further configured to:
classify, using a third machine learning model, the image to determine a third portion of the identification data, wherein
the second portion of the identification data comprises a general identifier for the image, and
the third portion of the identification data comprises an indication of a feature in the image.
15. The system of claim 14 , wherein
the individual identifier identifies a vehicle listing and the data set is a vehicle listing data set;
the member of the data set corresponds to the feature in the image, and
the generation of the mismatch alert is in response to a mismatch between the third portion comprising the indication of the feature and the member of the data set corresponding to the feature.
16. The system of claim 10 , wherein the processor and the memory are further configured to:
classify, using an alternate machine learning model, the image to determine a third portion of the identification data, in response to a failure to classify the image using the first machine learning model, wherein the third portion of the identification data comprises a second general image class; and
generate a classification failure alert, in response to failures to classify the image using the first machine learning model and the alternate machine learning model, wherein the classification failure alert includes a notification that the image is not a vehicle image.
17. The system of claim 10 , wherein the processor and the memory are further configured to modify the image from the data set, in response to the mismatch between the at least one portion of the identification data and the member of the data set.
18. The system of claim 10 , wherein the processor and the memory are further configured to modify the member of the data set in response to the mismatch between the at least one portion of the identification data and the member of the data set.
19. The system of claim 10 , wherein the processor and the memory are further configured to:
determine a first modification, among two or more candidate modifications, the first modification comprising modifying the member of the data set;
determine a second modification, among the two or more candidate modifications, the second modification comprising replacing the image to change the identification data;
for each of the first modification and the second modification, determine a number of steps required to eliminate mismatches between the identification data and all members of the data set;
perform a modification, among the two or more candidate modifications, having least number of steps.
20. A non-transitory computer-readable device having instructions stored thereon that, when executed by at least one computing device, cause the at least one computing device to perform operations comprising:
retrieving an image from a data set corresponding to an individual identifier;
classifying, using a first machine learning model, the image to determine a first portion of identification data for the image;
classifying, using a second machine learning model, the image to determine a second portion of the identification data, in response to successfully classifying the image using the first machine learning model;
comparing at least one portion of the identification data to a member of the data set to determine whether there is a mismatch between the at least one portion of the identification data and the member of the data set; and
generating a mismatch alert to indicate an error in the data set, in response to the mismatch between the at least one portion of the identification data and the member of the data set.