Methods and apparatuses for amount of object using two dimensional image
AssigneeNuvi Labs Co., Ltd.
InventorsDae Hoon Kim, Jey Yoon Ru, Seung Woo Ji
Life of this patent
Estimating the term…
Abstract
Provided are a method and an apparatus for measuring an object quantity using a 2D image. A method for measuring an object quantity using a 2D image according to one embodiment of the present disclosure comprises learning an object region model in a first object image which is a pre-learning target, learning an object quantity model for a first object region in the first object image using a feature map extracted from the first object image, and measuring an object quantity of at least one of a second object region and a background region in a second object image which is a measurement target using the learned object region model and object quantity model.
Claims12
A method for accurately measuring an object quantity using a 2D image executed by an apparatus for measuring the object quantity, the method comprising:
- learning an object region model in a first object image which is 2D image and a pre-learning target;
- learning an object quantity model for a first object region in the first object image using a feature map extracted from the first object image;
- and
- measuring an object quantity of at least one of a second object region and a background region in a second object image which is 2D image and a measurement target using the learned object region model and object quantity model,
- wherein the learning an object quantity model learns an object quantity model for the first object region by extracting a first feature map for a C-level part in order of object size in the first object image through an object quantity branch connected to the object region model, extracting a second feature map for a P-level part by mutually complementing the extracted first feature map for the C-level part, combining the extracted second feature maps, and passing the combined feature map to a fully connected layer, wherein the C-level part has feature point information in the order of the object size in the first object image, wherein the second feature map for the P-level part includes object quantity information,
- wherein the learning an object region model learns the object region model by segmenting an object of the first object region through classification of an object and a background in the first object region of the first object image.
An apparatus for measuring an object quantity using a 2D image, the apparatus comprising:
- a database storing a first object image which is a pre-learning target;
- a memory storing one or more programs;
- and
- a processor executing the stored one or more programs, wherein the processor is configured to:
- learn an object region model in a first object image which is a pre-learning target;
- learn an object quantity model for a first object region in the first object image using a feature map extracted from the first object image;
- and
- measure an object quantity of at least one of a second object region and a background region in a second object image which is a measurement target using the learned object region model and object quantity model,
- wherein the processor learns an object quantity model for the first object region by extracting a first feature map for a C-level part in order of object size in the first object image through an object quantity branch connected to the object region model, extracting a second feature map for a P-level part by mutually complementing the extracted first feature map for the C-level part, combining the extracted second feature maps, and passing the combined feature map to a fully connected layer, wherein the C-level part has feature point information in the order of the object size in the first object image, wherein the second feature map for the P-level part includes object quantity information,
- wherein the learning an object region model learns the object region model by segmenting an object of the first object region through classification of an object and a background in the first object region of the first object image.
Dependencies are inferred from the claim text.