US 12136276

Summary of U.S. Patent 12,136,276

A review of U.S. Patent 12,136,276 reveals a technology focused on the automatic calibration of cameras used in vehicles for lane detection and distance estimation. The patent is the subject of a legal challenge filed in 2026.

Title: Camera initialization for lane detection and distance estimation using single-view geometry

Assignee: Motive Technologies Inc.

Inventors: Ali Hassan, Ahmed Ali, Syed Wajahat Ali Shah Kazmi

Filing Date: December 1, 2023

Issue Date: November 5, 2024

Abstract:
The patent discloses methods, devices, and computer-readable media for detecting lanes and objects in image frames from a monocular camera. One embodiment describes a method that involves receiving a plurality of images and identifying a horizon by inputting the images into a deep learning (DL) model, which can be stored locally or accessed via a network. Based on the identified horizon, one or more camera parameters are determined and then stored or used to initialize the camera.

Overview of Independent Claims

The patent includes three independent claims, which outline the core inventive concepts.

Independent Claim 1: This claim describes a computer-implemented method for initializing a camera within a vehicle. The method involves:

1. Receiving video frames from the in-vehicle camera.
2. Using a convolutional neural network (CNN) to detect a horizon line and at least two lane lines within the video frames.
3. Calculating camera parameters, including the camera's height and a road plane normal vector, based on the detected horizon and lane lines.
4. Storing these calculated camera parameters for subsequent use.

In essence, this claim covers an automated process that uses AI to analyze video, identify key features of the road scene (horizon and lanes), and then mathematically determine the camera's position and orientation relative to the road, saving this information for later use in driver-assistance or autonomous driving systems.

Independent Claim 9: This claim focuses on a device, specifically an on-vehicle system, designed to perform camera initialization. The device consists of:

1. A memory for storing instructions.
2. One or more processors that execute these instructions to perform a series of actions.
   These actions are identical to the steps outlined in Claim 1: receiving video, using a CNN to detect the horizon and lane lines, calculating the camera's height and road plane normal, and storing these parameters.

This claim protects the physical hardware system that is configured to carry out the automated camera initialization method described in the first claim.

Independent Claim 17: This claim covers a non-transitory computer-readable storage medium. This medium contains stored instructions that, when executed by one or more processors, cause the system to perform the same method as detailed in Claim 1. The steps include:

1. Receiving video frames from the in-vehicle camera.
2. Detecting a horizon line and at least two lane lines using a CNN.
3. Calculating the camera's height and road plane normal from the detected lines.
4. Storing the calculated camera parameters.

This claim protects the software itself—the set of instructions that enables a computing device to perform the patented method, regardless of the specific hardware it is run on.

Legal Status:
It is important to note that U.S. Patent 12,136,276 is the subject of an Inter Partes Review (IPR) proceeding (Case IPR2026-00034) before the Patent Trial and Appeal Board of the USPTO. An IPR is a trial proceeding to review the patentability of one or more claims of a patent. The outcome of this proceeding could affect the validity of the claims described above. Information regarding any related litigation in the Court of Appeals for the Federal Circuit (CAFC) for 2026 was not found.