# US 2026/0278810 A1 · Method, device, and system for tracking an object in a sequence of video frames

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

- Number: US 2026/0278810 A1
- Document ID: US20260278810A1
- Status: Published application
- Kind: A1
- Application number: 19/565,244
- Filed: 2026-03-12
- Published: 2026-09-17
- Earliest claimed: 2025-03-17
- Assignees: Axis AB
- Applicants: Axis AB
- Inventors: Niclas DANIELSSON; Xing DANIELSSON FAN
- Web page: https://patentgrep.com/patent/US20260278810A1

## CPC

- G06T 7/246: using feature-based methods, e.g. the tracking of corners or segments
- G06T 7/248: {involving reference images or patches}
- G06T 7/277: involving stochastic approaches, e.g. using Kalman filters
- G06T 7/74: {involving reference images or patches}
- G06T 2207/10016: Video; Image sequence
- G06T 2207/30232: Surveillance
- G06T 2207/30241: Trajectory
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

A method, device, and system track an object in a sequence of video frames by using a tracking filter. In the method, costs are assigned for associating each object detection in a received set of object detections to the tracked object. The cost assigned to an object detection is a measure of deviation between a motion associated with the object detection and both of a first motion of the object estimated by the tracking filter and a second motion of the object indicated by motion vectors from a video encoder which encodes the sequence of video frames.

## Claims

### Claim 1

A method for tracking an object in a sequence of video frames, comprising:
tracking an object in a sequence of video frames by using a tracking filter, wherein the tracking filter estimates a position and a first motion of the object in a current video frame,
receiving motion vectors from a video encoder which encodes the sequence of video frames, wherein the motion vectors are indicative of a second motion of the object in the current video frame,
receiving a set of object detections in a subsequent video frame in the sequence, wherein each object detection is associated with a motion that corresponds to a displacement between the estimated position of the object in the current video frame and a position of the object detection in the subsequent video frame,
assigning, for each object detection, a cost for associating the object detection to the tracked object, wherein the cost includes a measure of deviation between the motion associated with the object detection and both of the first motion in the current video frame estimated by the tracking filter and the second motion in the current video frame indicated by the motion vectors,
associating one of the object detections to the tracked object based on the costs assigned to the object detections, and
using the tracking filter to estimate a position and a first motion of the object in the subsequent video frame based on the associated object detection.

### Claim 2 (depends on 1)

The method of claim 1, wherein an impact of the second motion on the cost depends on a reliability level of the received motion vectors such that the impact increases with increasing reliability level.

### Claim 3 (depends on 2)

The method of claim 2, wherein the reliability level of the received motion vectors is determined from a signal to noise ratio in the current video frame, or
wherein the estimated position of the object in the current video frame corresponds to an object area in the current video frame, and the reliability level of the received motion vectors is determined from a level of coherency of the motion vectors in the object area.

### Claim 4 (depends on 1)

The method of claim 1, wherein the cost includes a combination of a first cost and a second cost, wherein the first cost is a measure of deviation between the motion associated with the object detection and the first motion in the current video frame estimated by the tracking filter, and the second cost is a measure of deviation between the motion associated with the object detection and the second motion in the current video frame indicated by the motion vectors.

### Claim 5 (depends on 4)

The method of claim 4, wherein the combination is a multiplicative combination or an additive combination of the first cost and the second cost.

### Claim 6 (depends on 4)

The method of claim 4,
wherein the tracking filter uses the estimated position and first motion of the object in the current video frame to predict a first position of the object in the subsequent video frame,
wherein the first cost is calculated as a positional deviation between the position of the detected object in the subsequent frame and the predicted first position of the object in the subsequent video frame.

### Claim 7 (depends on 4)

The method of claim 4, wherein the second cost depends on an angular deviation between a direction of the motion associated with the object detection and a direction of the second motion in the current video frame indicated by the motion vectors.

### Claim 8 (depends on 4)

The method of claim 4, further comprising:
predicting a second position of the object in the subsequent frame from the estimated position of the object in the current video frame and the second motion in the current video frame indicated by the motion vectors,
wherein the second cost is calculated as a positional deviation between the position of the detected object in the subsequent video frame and the predicted second position of the object in the subsequent video frame.

### Claim 9 (depends on 1)

The method of claim 1,
wherein the tracking filter uses the estimated position and first motion of the object in the current video frame to predict a first position of the object in the subsequent video frame,
wherein the method further comprises predicting a second position of the object in the subsequent video frame from the estimated position of the object in the current video frame and the second motion in the current video frame indicated by the motion vectors, and
wherein the cost is a positional deviation between the position of the detected object in the subsequent video frame and a combination of the predicted first and second positions in the subsequent video frame.

### Claim 10 (depends on 9)

The method of claim 9, wherein the combination of the predicted first and second positions in the subsequent video frame is a weighted average of the predicted first and second positions.

### Claim 11 (depends on 1)

The method of claim 1, wherein the method is applied to track a plurality of objects in the sequence of video frames,
wherein, in the step of associating, object detections are associated to the tracked plurality of objects so as to minimize a total cost of the object detections that are associated with the tracked plurality of objects.

### Claim 12 (depends on 1)

The method of claim 1, wherein the tracking filter estimates the position and the first motion of the object in the subsequent video frame by:
predicting a position and a first motion of the object in the subsequent video frame from the position and the first motion of the object in the current video frame, and
updating the predicted position and first motion of the object in the subsequent video frame in view of the associated object detection.

### Claim 13

A device for tracking an object in a sequence of video frames, comprising circuitry configured to:
track an object in a sequence of video frames by using a tracking filter, wherein the tracking filter estimates a position and a first motion of the object in a current video frame,
receive motion vectors from a video encoder which encodes the sequence of video frames, wherein the motion vectors are indicative of a second motion of the object in the current video frame,
receive a set of object detections in a subsequent video frame in the sequence, wherein each object detection is associated with a motion that corresponds to a displacement between the estimated position of the object in the current video frame and a position of the object detection in the subsequent video frame,
assign, for each object detection, a cost for associating the object detection to the tracked object, wherein the cost includes a measure of deviation between the motion associated with the object detection and both of the first motion in the current video frame estimated by the tracking filter and the second motion in the current video frame indicated by the motion vectors,
associate one of the object detections to the object track based on the costs assigned to the object detections, and
use the tracking filter to estimate a position and a first motion of the object in the subsequent video frame based on the associated object detection.

### Claim 14 (depends on 13)

A system for tracking an object in a sequence of video frames, comprising:
a video encoder configured to encode the sequence of video frames and produce motion vectors indicative of a motion of the object in the video frames
an object detector configured to detect objects in the video frames of the sequence of video frames, and
a device for tracking an object in the sequence of video frames according to claim 13, wherein the device receives motion vectors from the video encoder and object detections from the object detector.

### Claim 15

A non-transitory computer-readable medium comprising computer program code which, when executed by a device with processing capability, causes the device to carry out a method for tracking an object in a sequence of video frames, comprising:
tracking an object in a sequence of video frames by using a tracking filter, wherein the tracking filter estimates a position and a first motion of the object in a current video frame,
receiving motion vectors from a video encoder which encodes the sequence of video frames, wherein the motion vectors are indicative of a second motion of the object in the current video frame,
receiving a set of object detections in a subsequent video frame in the sequence, wherein each object detection is associated with a motion that corresponds to a displacement between the estimated position of the object in the current video frame and a position of the object detection in the subsequent video frame,
assigning, for each object detection, a cost for associating the object detection to the tracked object, wherein the cost includes a measure of deviation between the motion associated with the object detection and both of the first motion in the current video frame estimated by the tracking filter and the second motion in the current video frame indicated by the motion vectors,
associating one of the object detections to the tracked object based on the costs assigned to the object detections, and
using the tracking filter to estimate a position and a first motion of the object in the subsequent video frame based on the associated object detection.
