# US 2019/0244365 A1 · Systems and methods for predictive/reconstructive visual object tracker

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0244365 A1
- Document ID: US20190244365A1
- Status: Published application
- Kind: A1
- Application number: 16/357,536
- Filed: 2019-03-19
- Published: 2019-08-08
- Earliest claimed: 2016-06-17
- Assignees: Brain Corporation
- Applicants: Brain Corporation
- Inventors: Filip Piekniewski; Micah Richert; Dimitry Fisher; Patryk Laurent; Csaba Petre
- Web page: https://patentgrep.com/patent/US20190244365A1

## CPC

- G06T 7/20: Analysis of motion (motion estimation for coding, decoding, compressing or decompressing digital video signals H04N19/43, H04N19/51)
- G06T 1/60: Memory management
- G06T 7/0002: {Inspection of images, e.g. flaw detection}
- G06T 9/00: Image coding (bandwidth or redundancy reduction for static pictures H04N1/41; coding or decoding of static colour picture signals H04N1/64; methods or arrangements for coding, decoding, compressing or decompressing digital video signals H04N19/00)
- G06T 9/002: {using neural networks}
- G06T 9/004: {Predictors, e.g. intraframe, interframe coding}
- H04N 19/503: involving temporal prediction (adaptive coding with adaptive selection between spatial and temporal predictive coding H04N19/107; adaptive coding with adaptive selection among a plurality of temporal predictive coding modes H04N19/109)
- G06F 12/08: in hierarchically structured memory systems, e.g. virtual memory systems
- G06F 2212/1016: Performance improvement
- G06F 2212/401: Compressed data
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Systems and methods for predictive/reconstructive visual object tracking are disclosed. The visual object tracking has advanced abilities to track objects in scenes, which can have a variety of applications as discussed in this disclosure. In some exemplary implementations, a visual system can comprise a plurality of associative memory units, wherein each associative memory unit has a plurality of layers. The associative memory units can be communicatively coupled to each other in a hierarchical structure, wherein data in associative memory units in higher levels of the hierarchical structure are more abstract than lower associative memory units. The associative memory units can communicate to one another supplying contextual data.

## Claims

### Claim at position 1

-22. (canceled)

### Claim 23

A system for tracking an object, comprising:
an autoencoder configured to execute computer readable instructions to,
associate a present signal with a reconstructed present signal, the present signal being received by the autoencoder, the present signal includes information regarding the object, and
reconstruct the present signal based on at least the present signal and the association;
a predictive encoder configured to execute the computer readable instructions to,
associate a past signal with the present signal, the past and present signals being received by the predictive encoder, the past and present signals includes information regarding the object and a context,
predict a future signal based on at least the present signal, the association and the context,
compress the prediction of the future signal; and
associative memory units configured to receive the compressed prediction of the future signal by the predictive encoder.

### Claim 24 (depends on 23)

The system of claim 23, wherein predictive encoder is further configured to execute the computer readable instructions to construct a predicted data signal for a future time period prior to the associating of the past signal with the present signal, the predicted data signal corresponds to a motion of the object.

### Claim 25 (depends on 23)

The system of claim 23, wherein the associate memory units are configured to,
receive a plurality of input frames corresponding to data signals pertaining to the object at a first time and a second time, the second time is after the first time, and
associate the plurality of input frames with a plurality of output frames, the plurality of output frames being substantially similar to the plurality of input frames.

### Claim 26 (depends on 23)

The system of claim 23, wherein the associative memory units includes lower layer for inputs, a middle layer for processing, and an upper layer as the output, the lower, middle and upper layers being separate and distinct from each other.

### Claim 27 (depends on 26)

The system of claim 26, wherein the middle layer of the associate memory units is configured to receive the context and tracker readout, the context corresponds to spatial and/or temporal context of the object, and the tracker readout corresponds to presence and/or location of the object, the presence and/or location of the object corresponds to a heatmap.

### Claim 28 (depends on 23)

The system of claim 23, wherein the associative memory units is configured to execute the computer readable instructions to operate in a first configuration and a second configuration, the first configuration corresponds to the associative memory units creating associations between the past and present signals, the second configuration corresponds to making a prediction based on the present signal.

### Claim 29 (depends on 23)

The system of claim 23, wherein the associative memory units is configured to execute the computer readable instructions to provide an estimate as to whether the object is in view of an autonomous device.

### Claim 30

A method for tracking an object, comprising:
associating a present signal with a reconstructed present signal, the present signal being received by an autoencoder, the present signal includes information regarding the object;
reconstructing the present signal based on at least the present signal and the association;
associating a past signal with the present signal, the past and present signals being received by a predictive encoder, the past and present signals includes information regarding the object and a context;
predict a future signal based on at least the present signal, the association and the context;
compress the prediction of the future signal; and
transmit the compressed prediction of the future signal by the predictive encoder to associative memory units.

### Claim 31

A non-transitory computer readable medium having computer readable instructions stored thereon, that when executed by at least one processor, configure the at least one processor to,
associate a present signal with a reconstructed present signal, the present signal being received by an autoencoder, the present signal includes information regarding the object;
reconstruct the present signal based on at least the present signal and the association;
associate a past signal with the present signal, the past and present signals being received by a predictive encoder, the past and present signals includes information regarding the object and a context;
predict a future signal based on at least the present signal, the association and the context;
compress the prediction of the future signal; and
transmit the compressed prediction of the future signal by the predictive encoder to associative memory units.
