Apparatus and method for generating depth map by using volumetric feature
Assignee42Dot Inc.
InventorsSeong Gyun JEONG, Jung Hee KIM, Phuoc Tien Nguyen, Jun Hwa HUR
Life of this patent
Estimating the term…
Abstract
Provided are an apparatus and method for generating a depth map by using a volumetric feature. The method may generate a single feature map for a base image included in a surround-view image by performing encoding and postprocessing on the base image, and generate a volumetric feature by encoding the single feature map with depth information and then projecting a result of the encoding into a three-dimensional space. In this method, a depth map of a surround-view image may be generated by using a depth decoder to decode a volumetric feature.
Claims10
A method of obtaining pose information of a monocular camera by using a volumetric feature, the method comprising:
- obtaining a surround-view image through one or more monocular cameras;
- generating multi-scale image features by encoding a base image included in the surround-view image;
- generating a single feature map for the base image by resizing each of the multi-scale image features based on a preset resolution and concatenating the resized multi-scale image features with each other;
- generating a volumetric feature by encoding the single feature map with depth information and then projecting a result of the encoding into a three-dimensional space;
- and
- collectively obtaining pose information of the monocular camera at a next time point by using a pose decoder to decode the volumetric feature.
An apparatus for obtaining pose information of a monocular camera by using a volumetric feature the apparatus, comprising:
- a memory storing at least one program;
- and
- a processor configured to execute the at least one program to obtain a surround-view image through one or more monocular cameras, generate multi-scale image features by encoding a base image included in the surround-view image, generate a single feature map for the base image by resizing each of the multi-scale image features based on a preset resolution and concatenating the resized multi-scale image features with each other, generate a volumetric feature by encoding the single feature map with depth information and then projecting a result of the encoding into a three-dimensional space, and collectively obtain pose information of the monocular camera at a next time point by using a pose decoder to decode the volumetric feature.
Dependencies are inferred from the claim text.