18416982. OBJECT DETECTION DEVICE, OBJECT DETECTION METHOD, AND RECORDING MEDIUM simplified abstract (NEC Corporation)

From WikiPatents
Jump to navigation Jump to search

OBJECT DETECTION DEVICE, OBJECT DETECTION METHOD, AND RECORDING MEDIUM

Organization Name

NEC Corporation

Inventor(s)

Yusuke Takahashi of Tokyo (JP)

OBJECT DETECTION DEVICE, OBJECT DETECTION METHOD, AND RECORDING MEDIUM - A simplified explanation of the abstract

This abstract first appeared for US patent application 18416982 titled 'OBJECT DETECTION DEVICE, OBJECT DETECTION METHOD, AND RECORDING MEDIUM

Simplified Explanation

The abstract describes an object detection device that efficiently generates high-quality training data. The device includes a detection unit, a reception unit, a generation unit, and a learning unit.

  • Detection unit uses a dictionary to detect objects in an input image.
  • Reception unit displays the input image with emphasized partial areas of detected objects and receives user input selecting a partial area and its class.
  • Generation unit creates training data from the selected partial area image and inputted class.
  • Learning unit uses the training data to improve the object detection dictionary.

Potential Applications

This technology can be applied in various fields such as autonomous driving, surveillance systems, robotics, and image recognition software.

Problems Solved

1. Efficient generation of high-quality training data for object detection. 2. Improved accuracy and performance of object detection systems.

Benefits

1. Enhanced object detection capabilities. 2. Streamlined training data generation process. 3. Increased accuracy and reliability in object recognition tasks.

Potential Commercial Applications

"Enhancing Object Detection with Efficient Training Data Generation" can be utilized in industries such as security, retail, manufacturing, and healthcare for improved object recognition and tracking systems.

Possible Prior Art

Prior art may include similar object detection devices with training data generation capabilities, but this specific combination of features and functions may be unique to this patent application.

Unanswered Questions

How does the device handle complex or overlapping objects in the input image?

The abstract does not provide details on how the device deals with complex or overlapping objects during the detection process.

What is the computational efficiency of the device when generating training data from selected partial areas?

The abstract does not mention the computational efficiency or processing speed of the device during the training data generation phase.


Original Abstract Submitted

Provided is an object detection device or the like which efficiently generates good-quality training data. This object detection device is provided with: a detection unit which uses a dictionary to detect objects from an input image; a reception unit which displays, on a display device, the input image accompanied by a display emphasizing partial areas of detected objects, and receives, from one operation of an input device, a selection of a partial area and an input of the class of the selected partial area; a generation unit which generates training data from the image of the selected partial area and the inputted class; and a learning unit which uses the training data to learn the dictionary.