Method for identifying an object in an image

By enhancing images and then inputting them into an artificial neural network, the problems of object recognition accuracy and energy consumption in autonomous driving and drones are solved, achieving efficient object recognition results.

CN121789179APending Publication Date: 2026-04-03DEEPX CO LTD
View PDF 0 Cites -1 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2020-06-04
Publication Date
2026-04-03

AI Technical Summary

Technical Problem

Existing technologies struggle to accurately and in real-time identify objects in rapidly changing driving environments in autonomous driving and drones, and high-performance deep learning processors are difficult to apply due to power consumption and heat generation issues.

Method used

The received image is enhanced by performing image enhancement processing, including noise reduction, deblurring, high dynamic range, tone mapping, dehazing, brightness and contrast adjustment, automatic white balance, and decompression, and then the image quality is improved before being input into an artificial neural network for object recognition.

Benefits of technology

It improves object recognition rate, enabling high-precision object recognition in autonomous driving and drones, while reducing power consumption and heat generation issues.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121789179A_ABST
    Figure CN121789179A_ABST
Patent Text Reader

Abstract

An apparatus for recognizing an object in an image according to an embodiment of the present disclosure may include: a preprocessing module configured to receive an image including an object and output a preprocessed image by performing image enhancement processing on the received image, the recognition rate of the object included in the received image is improved; and an object recognition module configured to recognize the object included in the image by inputting the preprocessed image to an input layer of an artificial neural network for object recognition.
Need to check novelty before this filing date? Find Prior Art