Granted Patent B2

US 12,456,297 · App. 17/839,805 · Granted Oct 28, 2025

Inventors: Allen Atamer (Toronto, CA); Tom Lee (Toronto, CA)

Assignee: AVENU STR IP LLC

G06V20/176G06F16/5846G06V10/764

Abstract

Claims (50)

1. A method for determining a location of an unlocated short term rental (STR) using image recognition, comprising:

receiving an unlocated STR query, the unlocated STR query having an unlocated image associated with the unlocated STR;

analyzing the unlocated image with image recognition to assign an unlocated image archetype to the unlocated image;

retrieving a list of candidate locations from a candidate data source using the unlocated image, each of the candidate locations having an associated candidate location image, and the candidate location image being assigned a candidate location image archetype by analyzing the candidate location image;

comparing the unlocated image and the candidate location image in response to the archetype of the unlocated image being the same as the archetype of the candidate location image for generating a matching score;

determining the location of the unlocated STR by associating the candidate location with the unlocated image when the matching score is over a threshold.

2. The method of claim 1 , wherein the step of analyzing the unlocated image further comprises assigning the archetype of the unlocated image by classifying the unlocated image using a machine learning framework for image archetype classification.

3. The method of claim 1 , wherein the step of analyzing the unlocated image further comprises:

analyzing the unlocated image with an object detection model, the object detection model for applying an object recognition algorithm to detect one or more objects within the unlocated image; and

assigning an archetype to the unlocated image based on the detected objects within the unlocated image.

4. The method of claim 1 , wherein the unlocated STR query further contains an initial longitudinal and latitudinal search coordinate, and an area of search related to the initial search coordinate; and

further wherein the candidate locations retrieved are limited to locations within the area of search.

5. The method of claim 4 , further comprising retrieving an initial candidate location through a reverse geocoding application programming interface (API) and the initial search coordinate;

wherein the candidate locations retrieved from the data source also uses the initial candidate location.

6. The method of claim 1 , further comprising determining a dwelling unit type of the unlocated STR query from the archetype assigned to the unlocated image;

wherein the candidate locations retrieved are limited to the candidate locations with the same candidate dwelling unit type.

7. The method of claim 1 , further comprising extracting text and metadata from the unlocated image;

wherein the candidate locations retrieved from the data source also uses the extracted text and metadata.

8. The method of claim 1 , further comprising determining a number of floors for a building object in the unlocated image by using a Fourier transform to extract the period of repeating floors for the building object from the unlocated image;

wherein a floor or a height of the observation point of the unlocated image is determined from the number of floors for the building object in the unlocated image.

9. The method of claim 1 , further comprising:

identifying two or more landmark objects in the unlocated image; and

triangulating each of the candidate locations of the unlocated image from the two or more landmark objects in the unlocated image.

10. A system for determining a location of an unlocated short term rental (STR) using image recognition, comprising:

a comparison view component operative to provide a view and interface for comparing image information about an unlocated STR unit and to identify a location for the unlocated STR unit, wherein the comparison view component receives an unlocated STR query, the unlocated STR query having an unlocated image associated with the unlocated STR;

a photo classification component operative using image recognition to analyze the unlocated image to assign an unlocated image archetype to the unlocated image;

a reverse photo search component to retrieve images and;

wherein the system is operative to retrieve a list of candidate locations from a candidate data source using the unlocated image, each of the candidate locations having an associated candidate location image, and the candidate location image being assigned a candidate location image archetype by analyzing the candidate location image;

wherein the system is further operative to compare the unlocated image and the candidate location image in response to the archetype of the unlocated image being the same as the archetype of the candidate location image for generating a matching score; and

wherein the system is operative to determine a location of the unlocated STR by associating the candidate location with the unlocated image when the matching score is over a threshold.

11. The system of claim 10 , wherein the photo classification component is further operative to assign the archetype of the unlocated image by classifying the unlocated image using a machine learning framework for image archetype classification.

12. The system of claim 10 , wherein the photo classification component further comprises an object detection model for applying an object recognition algorithm to detect one or more objects within the unlocated image and assigning an archetype to the unlocated image based on the detected objects within the image.

13. The system of claim 10 , wherein the unlocated STR query further contains an initial longitudinal and latitudinal search coordinate, and an area of search related to the initial search coordinate; and

further wherein the candidate locations retrieved are limited to locations within the area of search.

14. The system of claim 13 , further comprising retrieving an initial candidate location through a reverse geocoding application programming interface (API) and the initial search coordinate;

wherein the candidate locations retrieved from the data source also uses the initial candidate location.

15. The system of claim 10 , further comprising determining a dwelling unit type of the unlocated STR query from the archetype assigned to the unlocated image;

wherein the candidate locations retrieved are limited to the candidate locations with the same candidate dwelling unit type.

16. The system of claim 10 , further comprising a metadata extraction component operative for assigning location information to unlocated images;

wherein the system extracts text and metadata from the unlocated image; and the retrieving of the candidate locations from the data source also uses the extracted text and metadata.

17. The system of claim 10 , wherein the floor counting component is further operative to determine a number of floors for a building object in the unlocated image using a Fourier transform to extract the period of repeating floors for the building object from the unlocated image;

wherein a floor or a height of the observation point of the unlocated image is determined from the number of floors for the building object in the unlocated image.

18. The system of claim 10 , further comprising a landmarks triangulation component operative to identify two or more landmark objects in the unlocated image and to triangulate each of the candidate locations of the observation point of the unlocated image from the two or more landmark objects in the unlocated image.

19. A method for determining a location of an unlocated short term rental (STR) using image recognition, comprising:

receiving an unlocated STR query, the unlocated STR query having unlocated images associated with the unlocated STR;

analyzing the unlocated images using image recognition to assign an unlocated image archetype to each of the unlocated images;

retrieving a list of candidate locations from a candidate data source using the unlocated images, each of the candidate locations in the list having an associated candidate location image, and each of the candidate location images being assigned a candidate location image archetype by analyzing each of the candidate location images;

comparing the unlocated images and the candidate location images in response to the archetype of one of the unlocated images being the same as the archetype of one of the candidate location images for generating a total matching score for each of the candidate locations in the list; and

presenting through a user interface the list of candidate locations sorted by the total matching score for each of the candidate locations in the list.

20. The method of claim 19 , further wherein the total matching score for each of the candidate locations is generated by adding together a matching score from the comparison between each of the unlocated images with each of the candidate images with the same image archetype.

Assignments (7)

Continuity (2)

References Cited (20)