Method for providing object detecting system capable of updating types of detectable classes in real-time by using continual learning and device using the same
A technology of object detection and real-time update, applied in neural learning methods, still image data clustering/classification, scene recognition, etc., can solve expensive and other problems
Active Publication Date: 2020-08-07
STRADVISION
View PDF12 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, the learning process of updating the parameters of the CNN is expensive, and the process of transmitting the parameters and related information to each autonomous vehicle for updating is also expensive
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment approach
[0041] In order to enable those skilled in the art to easily implement the present disclosure, exemplary embodiments of the present invention will be described in detail by referring to the accompanying drawings, as shown below.
[0042] figure 1 is a diagram schematically showing a configuration of an object detection system performing a method for providing an object detection system capable of updating categories of detectable classes in real time according to an exemplary embodiment of the present invention.
[0043] by reference figure 1 , the object detection system 100 may include a recognizer 110 , a characteristic fingerprint database 120 , an unknown image database 130 and a known image database 140 described later. Input / output processing and calculation of the recognizer 110, the feature fingerprint database 120, the unknown image database 130, and the known image database 140 may be performed by a communication section and a processor included in the object detec...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
A method for updating an object detecting system to detect objects with untrained classes in real-time is provided. The method includes steps of: (a) the object detecting system, if at least one inputimage is acquired, instructing a recognizer included therein to generate a specific feature map, and to generate a specific query vector; (b) the object detecting system instructing a similarity determining unit (i) to compare the specific query vector to data vectors, to thereby calculate each of first similarity scores between the specific query vector and each of the data vectors, and (ii) toadd a specific partial image to an unknown image DB, if a specific first similarity score is smaller than a first threshold value; and (c) the object detecting system, if specific class information isacquired, instructing a short-term update unit to generate a specific short-term update vector, and update the feature fingerprint DB.
Description
technical field [0001] The present invention relates to a method and apparatus for self-driving vehicles; more particularly, to a method and apparatus for providing an object detection system capable of updating the categories of detectable classes in real time. Background technique [0002] Deep Convolution Neural Networks (Deep CNNs) are at the heart of remarkable developments in deep learning. CNNs were already used to solve character recognition problems in the 1990s, but thanks to recent research, their use has become as widespread as it is now. In the 2012 ImageNet image classification competition, CNN defeated other competitors to win. Afterwards, convolutional neural networks became a very useful tool in the field of Machine Learning. [0003] Recently, CNNs have been widely used in autonomous driving. CNN can perform various functions to support the automatic driving of self-driving cars, such as Semantic Segmentation, object detection, etc., these functions are ...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/62G06F16/583G06F16/50G06V10/764G06V10/774
CPCG06F16/583G06F16/50G06V20/56G06F18/22G06F18/214G06N3/08G06F16/55G06F16/5854G06V10/96G06V10/462G06V2201/08G06V10/82G06V10/764G06V10/774G06N3/045G06N3/082G06V40/1335G06V40/1347G06V40/1365G06N3/04G06V20/58G06V2201/07G06F18/2155G06F18/213
Inventor 金桂贤金镕重金鹤京南云铉夫硕熏成明哲申东洙吕东勋柳宇宙李明春李炯树张泰雄郑景中诸泓模赵浩辰
Owner STRADVISION