18451287. METHOD AND APPARATUS WITH OBJECT DETECTOR TRAINING simplified abstract (Samsung Electronics Co., Ltd.)

From WikiPatents
Jump to navigation Jump to search

METHOD AND APPARATUS WITH OBJECT DETECTOR TRAINING

Organization Name

Samsung Electronics Co., Ltd.

Inventor(s)

Sujin Jang of Suwon-si (KR)

Sangpil Kim of Seoul (KR)

Jinkyu Kim of Seoul (KR)

Wonseok Roh of Seoul (KR)

Gyusam Chang of Seongnam-si (KR)

Dongwook Lee of Suwon-si (KR)

Dae Hyun Ji of Suwon-si (KR)

METHOD AND APPARATUS WITH OBJECT DETECTOR TRAINING - A simplified explanation of the abstract

This abstract first appeared for US patent application 18451287 titled 'METHOD AND APPARATUS WITH OBJECT DETECTOR TRAINING

Simplified Explanation

The abstract describes a method and apparatus for object detector training, involving obtaining input data from a target object, performing data augmentation, extracting features to a shared embedding space, identifying loss functions, and updating weights based on the loss functions.

  • Obtaining input data from a target object
  • Performing data augmentation on the input data
  • Extracting features to a shared embedding space
  • Identifying loss functions based on the extracted features
  • Updating weights of the encoder based on the loss functions

Potential Applications

This technology can be applied in various fields such as computer vision, autonomous driving, surveillance systems, and robotics for object detection and recognition tasks.

Problems Solved

This technology helps improve the accuracy and efficiency of object detection systems by training the detectors with augmented data and shared embedding spaces, leading to better feature extraction and loss function identification.

Benefits

The benefits of this technology include enhanced object detection performance, increased robustness to variations in input data, and improved generalization capabilities for different object detection tasks.

Potential Commercial Applications

The potential commercial applications of this technology include developing advanced object detection systems for security, retail, manufacturing, and healthcare industries, as well as for autonomous vehicles and drones.

Possible Prior Art

One possible prior art for this technology could be the use of data augmentation techniques in machine learning and computer vision to improve model performance and generalization capabilities.

Unanswered Questions

How does this technology compare to existing object detection methods?

This article does not provide a direct comparison with existing object detection methods in terms of performance, efficiency, or accuracy.

What are the computational requirements for implementing this technology?

The article does not mention the computational resources needed to implement this method, such as processing power, memory, or training time.


Original Abstract Submitted

A method and apparatus with object detector training is provided. The method includes obtaining first input data and second input data from a target object; obtaining second additional input data by performing data augmentation on the second input data; extracting a first feature to a shared embedding space by inputting the first input data to a first encoder; extracting a second feature to the shared embedding space by inputting the second input data to a second encoder; extracting a second additional feature to the shared embedding space by inputting thesecond additional input data to the second encoder; identifying a first loss function based on the first feature, the second feature, and the second additional feature; identifying a second loss function based on the second feature and the second additional feature; and updating a weight of the second encoder based on the first loss function and the second loss function.