Method and apparatus for predicting feature space decay using variational auto-encoder networks

Patent metadata

Field	Value
Publication number	US-11170251-B2
Application number	US-202016864489-A
Country	US
Kind code	B2
Filing date	May 1, 2020
Priority date	Oct 30, 2018
Publication date	Nov 9, 2021
Grant date	Nov 9, 2021

Abstract

An apparatus, method and computer program product are provided for predicting feature space decay using variational auto-encoder networks. Methods may include: receiving a first image of a road segment including a feature disposed along the road segment; applying a loss function to the feature of the first image; generating a revised image, where the revised image includes a weathered iteration of the feature; generating a predicted image using interpolation between the image and the revised image of a partially weathered iteration of the feature; receiving a user image, where the user image is received from a vehicle traveling along the road segment; correlating a feature in the user image to the partially weathered iteration of the feature in the predicted image; and establishing that the feature in the user image is the feature disposed along the road segment.

First claim

That which is claimed:

1. An apparatus comprising at least one processor and at least one non-transitory memory including computer program code instructions stored thereon, the computer program code instructions configured to, when executed, cause the apparatus to at least:
receive a first image including a first feature;
apply a loss function to the first feature of the first image;
generate a revised image based on the application of the loss function to the first feature, wherein the revised image comprises a weathered iteration of the first feature; and
predict a partially-weathered iteration of the first feature based on the first image and the revised image;
wherein the partially-weathered iteration of the first feature permits the first feature to be identified in a received second image.

2. The apparatus of claim 1 , wherein the apparatus is further caused to:
receive the second image, wherein the second image comprises a partially-weathered first feature;
compare the partially-weathered first feature against the partially-weathered iteration of the first feature; and
establish an accuracy of the loss function.

3. The apparatus of claim 2 , wherein the apparatus is further caused to:
revise the partially-weathered iteration of the first feature to correlate to the partially weathered first feature in response to the partially-weathered first feature differing from the partially-weathered iteration of the first feature by more than a predefined amount.

4. The apparatus of claim 2 , wherein the apparatus is further caused to:
update the loss function in response to the partially-weathered first feature differing from the partially-weathered iteration of the first feature by more than a predefined amount.

5. The apparatus of claim 4 , wherein the loss function is updated to reflect the partially-weathered first feature of the second image.

6. The apparatus of claim 1 , wherein the loss function comprises a negative log-likelihood function.

7. The apparatus of claim 1 , wherein the apparatus is further caused to:
update the loss function using images of the first feature in various stages of decay.

8. The apparatus of claim 1 , wherein causing the apparatus to generate a revised image based on the application of the loss function to the first feature comprises causing the apparatus to:
generate a plurality of revised images by applying a number of weathered iterations to the first feature, the revised images representing a feature-decay space of the first feature.

9. The apparatus of claim 8 , wherein the apparatus is further caused to train a perception system using the plurality of revised images.

10. A computer program product comprising at least one non-transitory computer-readable storage medium having computer-executable program code instructions stored therein, the computer-executable program code instructions comprising program code instructions to:
receive a first image including a first feature;
apply a loss function to the first feature of the first image;
generate a revised image based on the application of the loss function to the first feature, wherein the revised image comprises a weathered iteration of the first feature; and
predict a partially-weathered iteration of the first feature based on the first image and the revised image;
wherein the partially-weathered iteration of the first feature permits the first feature to be identified in a received second image.

11. The computer program product of claim 10 , further comprising program code instructions to:
receive the second image, wherein the second image comprises a partially-weathered first feature;
compare the partially-weathered first feature against the partially-weathered iteration of the first feature; and
establish an accuracy of the loss function.

12. The computer program product of claim 11 , further comprising program code instructions to:
revise the partially-weathered iteration of the first feature to correlate to the partially weathered first feature in response to the partially-weathered first feature differing from the partially-weathered iteration of the first feature by more than a predefined amount.

13. The computer program product of claim 11 , further comprising program code instructions to:
update the loss function in response to the partially-weathered first feature differing from the partially-weathered iteration of the first feature by more than a predefined amount.

14. The computer program product of claim 13 , wherein the loss function is updated to reflect the partially-weathered first feature of the second image.

15. The computer program product of claim 10 , wherein the loss function comprises a negative log-likelihood function.

16. The computer program product of claim 10 , further comprising program code instructions to:
update the loss function using images of the first feature in various stages of decay.

17. The computer program product of claim 10 , wherein the program code instructions to generate a revised image based on the application of the loss function to the first feature comprises program code instructions to:
generate a plurality of revised images by applying a number of weathered iterations to the first feature, the revised images representing a feature-decay space of the first feature.

18. The computer program product of claim 17 , further comprising program code instructions to train a perception system using the plurality of revised images.

19. A method comprising:
receiving a first image including a first feature;
applying a loss function to the first feature of the first image;
generating a revised image based on the application of the loss function to the first feature, wherein the revised image comprises a weathered iteration of the first feature; and
predicting a partially-weathered iteration of the first feature based on the first image and the revised image;
wherein the partially-weathered iteration of the first feature permits the first feature to be identified in a received second image.

20. The method of claim 19 , further comprising:
receiving the second image, wherein the second image comprises a partially-weathered first feature;
comparing the partially-weathered first feature against the partially-weathered iteration of the first feature; and
establishing an accuracy of the loss function.

Assignees

• Here Global Bv

Inventors

• Viswanathan Anirudh

Classifications

• G01C21/26 Primary

  specially adapted for navigation in a road network · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/454

  Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN] · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06V20/588

  Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road · CPC title