18508904. OBJECT RECOGNITION APPARATUS FOR AN AUTONOMOUS VEHICLE AND AN OBJECT RECOGNITION METHOD THEREFOR simplified abstract (Kia Corporation)
Contents
OBJECT RECOGNITION APPARATUS FOR AN AUTONOMOUS VEHICLE AND AN OBJECT RECOGNITION METHOD THEREFOR
Organization Name
Inventor(s)
OBJECT RECOGNITION APPARATUS FOR AN AUTONOMOUS VEHICLE AND AN OBJECT RECOGNITION METHOD THEREFOR - A simplified explanation of the abstract
This abstract first appeared for US patent application 18508904 titled 'OBJECT RECOGNITION APPARATUS FOR AN AUTONOMOUS VEHICLE AND AN OBJECT RECOGNITION METHOD THEREFOR
Simplified Explanation: The patent application describes an object recognition apparatus and method for an autonomous vehicle. The apparatus includes a processor and storage that train network models based on images, correct distortions, generate distorted images, and perform fine-tuning for improved object recognition.
- Trains a first network model based on an image
- Corrects distortion and rotation of the image
- Generates at least one distorted image based on the corrected image
- Trains a feature transfer module based on features extracted from the corrected image and distorted image
- Inserts the feature transfer module into the first network model
- Performs fine-tuning for a second network model including the feature transfer module based on the distorted image
Potential Applications: 1. Autonomous vehicles for improved object recognition and navigation 2. Surveillance systems for enhanced security and threat detection 3. Robotics for better object manipulation and interaction
Problems Solved: 1. Addressing image distortion and rotation for accurate object recognition 2. Enhancing feature extraction and transfer for improved network training 3. Improving overall object recognition performance in autonomous systems
Benefits: 1. Increased accuracy in identifying objects in various environments 2. Enhanced efficiency in processing and analyzing images 3. Improved reliability and safety in autonomous vehicle operations
Commercial Applications: The technology can be applied in industries such as autonomous vehicles, security systems, and robotics for advanced object recognition capabilities, leading to safer and more efficient operations.
Questions about Object Recognition Technology: 1. How does the feature transfer module improve object recognition accuracy? 2. What are the potential limitations of this object recognition technology in real-world scenarios?
Frequently Updated Research: Ongoing research focuses on optimizing the feature transfer module for various object recognition tasks and exploring its applications in different industries.
Original Abstract Submitted
An object recognition apparatus and method for an autonomous vehicle are disclosed. The object recognition apparatus includes a processor and storage. The processor: trains a first network model based on an image; corrects distortion and rotation of the image; generates at least one distorted image based on the corrected image; trains a feature transfer module based on a feature extracted from the corrected image and a feature extracted from the at least one distorted image; inserts the feature transfer module into the first network model; and performs fine-tuning for a second network model including the feature transfer module based on the at least one distorted image.
