Transportation prediction system and method

• US 10,846,544 B2
• Filed: 07/16/2018
• Issued: 11/24/2020
• Est. Priority Date: 07/16/2018
• Status: Active Grant

• Current Assignee: Autobrains Technologies Ltd., Cortica Ltd.
• Original Assignee: Cartica Al Ltd.
• Inventors: Raichelgauz, Igal, Odinaev, Karina
• Primary Examiner(s): Neff, Michael R
• Application Number:
  US16/035,732
• Publication Number:
  US 20200019793A1
• Time in Patent Office: 862 Days
• Field of Search: 382104
• CPC Class Codes: B60W 30/0956 the prediction being responsive to traffic or environmental parameters
• CPC Class Codes: G06F 16/5838 using colour
• CPC Class Codes: G06V 20/30 in albums, collections or shared content, e.g. social network photos or video
• CPC Class Codes: G06V 20/58 Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads

Abstract

In one embodiment, an automotive prediction system includes a processing circuitry to obtain labels labelling media content elements identified in road-scene media content sequences, each label including a content descriptor selected from different content descriptors describing at least one media content element, the different content descriptors including a first and second content descriptor, calculate a correlation of the first and second content descriptor based on a count of occurrences of the first content descriptor being used for labelling after, but within a given temporal distance of the first content descriptor being used for labelling in the road-scene media content sequences, and populate an automotive prediction database with the correlation of the first and second content descriptor for use in making decisions during driving of a vehicle. Related apparatus and methods are also described.

Claims

1. An automotive prediction system comprising:

a processing circuitry; and

a memory to store data used by the processing circuitry, the processing circuitry being configured to;

obtain a plurality of labels labelling a plurality of media content elements identified in a plurality of road-scene media content sequences being at least one of audio or video road-scene sequences, each one label of the plurality of labels including a content descriptor selected from a plurality of different content descriptors describing at least one media content element of the plurality of media content elements, the plurality of different content descriptors including a first content descriptor and a second content descriptor;

calculate a correlation of the first content descriptor and the second content descriptor based on a first count of occurrences of the second content descriptor being used for labelling in the plurality of road-scene media content sequences after, but within a given temporal distance of, the first content descriptor being used for labelling in the plurality of road-scene media content sequences;

calculate a first correlation of the first content descriptor and the second content descriptor based on a second count of occurrences of the second content descriptor being used for labelling in the plurality of road-scene media content sequences after a first temporal distance from the first content descriptor being used for labelling in the plurality of road-scene media content sequences; and

calculate a second correlation of the first content descriptor and the second content descriptor based on a third count of occurrences of the second content descriptor being used for labelling in the plurality of road-scene media content sequences after a second temporal distance from the first content descriptor being used for labelling in the plurality of road-scene media content sequences, the second temporal distance being greater than the first temporal distance; and

populate an automotive prediction database with the correlation of the first content descriptor and the second content descriptor for use in making decisions during driving of a vehicle.

2. The system according to claim 1, wherein the processing circuitry is further configured to calculate a third correlation of the first content descriptor and the second content descriptor based on a third count of occurrences of the second content descriptor and the first content descriptor being used for labelling in the plurality of road-scene media content sequences at a same temporal location in the plurality of road-scene media content sequences.

3. The system according to claim 1, wherein each one label of at least some labels from the plurality of labels includes location data indicating a location of one media content element of the plurality of media content elements within a video frame of one road-scene media content sequence of the plurality of road-scene media content sequences, the processing circuitry being further configured to determine a location relationship between the first content descriptor and the second content descriptor based on the location data of the at least some labels.

4. The system according to claim 3, wherein the processing circuitry is further configured to determine the location relationship including trajectory prediction data of the second content descriptor based on a location or identified trajectory of the first content descriptor.

5. The system according to claim 1, wherein the processing circuitry is further configured to select the plurality of road-scene media content sequences from a second plurality of media content sequences based on selecting ones of the second plurality of media content sequences that include road-scene content descriptors.

6. The system according to claim 5, wherein the processing circuitry is further configured to select the plurality of road-scene media content sequences based on a geolocation origin of the second plurality of media content sequences.

7. The system according to claim 1, wherein the processing circuitry is further configured to improve any one or more of the following:

a speed and/or accuracy of calculating the correlation of the first content descriptor and the second content descriptor;

an accuracy of predictive correlation data for use in automotive prediction applications to predict a future road-scene event; and

an operation speed of a computing device calculating the correlation of the first content descriptor and the second content descriptor.

8. An automotive prediction system comprising:

a processing circuitry; and

a memory to store data used by the processing circuitry, the processing circuitry being configured to;

obtain a plurality of labels labelling a plurality of media content elements identified in a plurality of road-scene media content sequences being at least one of audio or video road-scene sequences, each one label of the plurality of labels including a content descriptor selected from a plurality of different content descriptors describing at least one media content element of the plurality of media content elements, the plurality of different content descriptors including a first content descriptor and a second content descriptor;

calculate a correlation of the first content descriptor and the second content descriptor based on a first count of occurrences of the second content descriptor being used for labelling in the plurality of road-scene media content sequences after, but within a given temporal distance of, the first content descriptor being used for labelling in the plurality of road-scene media content sequences; and

populate an automotive prediction database with the correlation of the first content descriptor and the second content descriptor for use in making decisions during driving of a vehicle;

determine a first value based on a second count of occurrences of the second content descriptor being used for labelling in the plurality of road-scene media content sequences;

determine a second value based on the first count of occurrences; and

calculate the correlation of the first content descriptor and the second content descriptor based on the first value and the second value.

9. The system according to claim 8, wherein the processing circuitry is further configured to determine the second value based on the first count of occurrences divided by a third count of occurrences of the first content descriptor being used for labelling in the plurality of road-scene media content sequences.

10. An automotive prediction system comprising:

a processing circuitry; and

a memory to store data used by the processing circuitry, the processing circuitry being configured to;

obtain a plurality of labels labelling a plurality of media content elements identified in a plurality of road-scene media content sequences being at least one of audio or video road-scene sequences, each one label of the plurality of labels including a content descriptor selected from a plurality of different content descriptors describing at least one media content element of the plurality of media content elements, the plurality of different content descriptors including a first content descriptor and a second content descriptor;

calculate a correlation of the first content descriptor and the second content descriptor based on a first count of occurrences of the second content descriptor being used for labelling in the plurality of road-scene media content sequences after, but within a given temporal distance of, the first content descriptor being used for labelling in the plurality of road-scene media content sequences; and

populate an automotive prediction database with the correlation of the first content descriptor and the second content descriptor for use in making decisions during driving of a vehicle;

wherein the plurality of content descriptors includes a third content descriptor which is different from the first content descriptor and the second content descriptor; and

the processing circuitry is further configured calculate a correlation of the first content descriptor, the second content descriptor, and the third descriptor based on a second count of occurrences of the third content descriptor being used for labelling in the plurality of road-scene media content sequences after a first temporal distance from the first content descriptor being used for labelling and after a second temporal distance from the second content descriptor being used for labelling in the plurality of road-scene media content sequences.

11. An automotive prediction method, comprising:

obtaining a plurality of labels labelling a plurality of media content elements identified in a plurality of road-scene media content sequences being audio and/or video road-scene sequences, each one label of the plurality of labels including a content descriptor selected from a plurality of different content descriptors describing at least one media content element of the plurality of media content elements, the plurality of different content descriptors including a first content descriptor and a second content descriptor;

calculating a correlation of the first content descriptor and the second content descriptor based on a count of occurrences of the second content descriptor and the first content descriptor being used for labelling within a given temporal distance of each other in the plurality of road-scene media content sequences;

calculating a first correlation of the first content descriptor and the second content descriptor based on a second count of occurrences of the second content descriptor being used for labelling in the plurality of road-scene media content sequences after a first temporal distance from the first content descriptor being used for labelling in the plurality of road-scene media content sequences;

calculating a second correlation of the first content descriptor and the second content descriptor based on a third count of occurrences of the second content descriptor being used for labelling in the plurality of road-scene media content sequences after a second temporal distance from the first content descriptor being used for labelling in the plurality of road-scene media content sequences, the second temporal distance being greater than the first temporal distance; and

populating an automotive prediction database with the correlation of the first content descriptor and the second content descriptor for use in making decisions during driving of a vehicle.

12. The method according to claim 11, further comprising calculating a third correlation of the first content descriptor and the second content descriptor based on a third count of occurrences of the second content descriptor and the first content descriptor being used for labelling in the plurality of road-scene media content sequences at a same temporal location in the plurality of road-scene media content sequences.