Image processing device and image processing method

The image processing apparatus and method address the challenge of accurately estimating lower end positions of multiple overlapping objects by employing contour analysis and time-series changes to separate and detect individual objects, enhancing accuracy in complex environments.

WO2026150739A1PCT designated stage Publication Date: 2026-07-16ASTEMO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
ASTEMO LTD
Filing Date
2025-12-16
Publication Date
2026-07-16

AI Technical Summary

Technical Problem

Existing image processing systems using monocular cameras struggle to accurately estimate the lower end positions of multiple overlapping objects due to misinterpretation of contour changes caused by both depth and vertical direction variations, leading to errors in distance calculation.

Method used

An image processing apparatus and method that includes a contour extraction unit, division determination unit, partial region generation unit, and object detection unit to correctly separate and detect individual objects by analyzing contour shapes and time-series changes, particularly in the vertical and horizontal directions, using techniques like semantic segmentation and neural networks.

Benefits of technology

Enables accurate estimation of lower end positions of each object, even in complex environments, reducing misestimation of objects with intricate shapes by effectively distinguishing and separating overlapping objects.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure JP2025043821_16072026_PF_FP_ABST
    Figure JP2025043821_16072026_PF_FP_ABST
Patent Text Reader

Abstract

Provided are an image processing device and an image processing method that make it possible to correctly estimate a lower end position for each of a plurality of objects in an environment in which the objects are mixed and to suppress misestimation for objects that have complicated shapes. The present invention comprises an outline extraction unit that extracts the outline of an object candidate region that is a region of an image in which an object has been captured, a division determination unit that, treating the direction in which a road surface has been captured in the image to be a first direction, determines, on the basis of the shape of the outline of the object candidate region at a first end in the first direction, whether a partial region that is a cut-out portion of the object candidate region is to be generated, a partial region generation unit that, when the division determination unit has determined that the partial region is to be generated, determines the position of the partial region on the basis of the shape of the outline at the first end and generates the partial region, and an object detection unit that performs object detection processing on the partial region.
Need to check novelty before this filing date? Find Prior Art