-
公开(公告)号：US20180365564A1
公开(公告)日：2018-12-20
申请号：US16004363
申请日：2018-06-09
Applicant: TuSimple
Inventor： Zehao HUANG , Naiyan WANG
Abstract: A method and device for training a neural network are disclosed. The method comprises: selecting, by a training device, a teacher network performing the same functions of a student network; and iteratively training the student network and obtaining a target network, through aligning distributions of features between a first middle layer and a second middle layer corresponding to the same training sample data, so as to transfer knowledge of features of a middle layer of the teacher network to the student network.
-
公开(公告)号：US20210256664A1
公开(公告)日：2021-08-19
申请号：US17308911
申请日：2021-05-05
Applicant: TUSIMPLE, INC.
Inventor： Pengfei CHEN , Nan YU , Naiyan WANG , Xiaodi HOU
Abstract: Disclosed are devices, systems and methods for processing an image. In one aspect a method includes receiving an image from a sensor array including an x-y array of pixels, each pixel in the x-y array of pixels having a value selected from one of three primary colors, based on a corresponding x-y value in a mask pattern. The method may further include generating a preprocessed image by performing preprocessing on the image. The method may further include performing perception on the preprocessed image to determine one or more outlines of physical objects.
-
Patent Agency Ranking
公开(公告)号：US20190279028A1
公开(公告)日：2019-09-12
申请号：US16273835
申请日：2019-02-12
Applicant: TuSimple
Inventor： Naiyan WANG , Jianfu ZHANG
Abstract: The present disclosure provides a method and an apparatus for object re-identification, capable of solving the problem in the related art associated with inefficiency and low accuracy of object re-identification based on multiple frames of images. The method includes, for each pair of objects: selecting one of a set of images associated with each of the pair of objects, to constitute a pair of current images for the pair of objects; inputting the pair of current images to a preconfigured feature extraction network, to obtain feature information for the pair of current images; determining whether the pair of objects are one and the same object based on the feature information for the pair of current images and feature information for one or more pairs of historical images for the pair of objects by using a preconfigured re-identification network; and outputting a determination result of said determining when the determination result is that the pair of objects are one and the same object or that the pair of objects are not one and the same object, or repeating the above steps using the pair of current images as a pair of historical images for the pair of objects when the determination result is that it is uncertain whether the pair of objects are one and the same object. With the solutions according to the present disclosure, the speed and efficiency of the object re-identification can be improved.
-
公开(公告)号：US20230351556A1
公开(公告)日：2023-11-02
申请号：US18344577
申请日：2023-06-29
Applicant: TUSIMPLE, INC. , BEIJING TUSEN WEILAI TECHNOLOGY CO., LTD.
Inventor： Pengfei CHEN , Nan YU , Naiyan WANG , Xiaodi HOU
CPC classification number: G06T5/002 , G06T5/009 , G06T5/004 , B60W40/02 , H04N25/13 , B60W2420/42 , G06T2207/10024
Abstract: Disclosed are devices, systems and methods for processing an image. In one aspect a method includes receiving an image from a sensor array including an x-y array of pixels, each pixel in the x-y array of pixels having a value selected from one of three primary colors, based on a corresponding x-y value in a mask pattern. The method may further include generating a preprocessed image by performing preprocessing on the image. The method may further include performing perception on the preprocessed image to determine one or more outlines of physical objects.
-
公开(公告)号：US20200020102A1
公开(公告)日：2020-01-16
申请号：US16577753
申请日：2019-09-20
Applicant: TUSIMPLE, INC.
Inventor： Hengchen DAI , Naiyan WANG
Abstract: The present disclosure provides a method and an apparatus for semantic segmentation of an image, capable of solving the problem in the related art associated with low speed and inefficiency in semantic segmentation of images. The method includes: receiving the image; performing semantic segmentation on the image to obtain an initial semantic segmentation result; and inputting image information containing the initial semantic segmentation result to a pre-trained convolutional neural network for semantic segmentation post-processing, so as to obtain a final semantic segmentation result. With the solutions of the present disclosure, the initial semantic segmentation result can be post-processed using the convolutional neural network, such that the speed and efficiency of the semantic segmentation of the image can be improved.
-
公开(公告)号：US20190318456A1
公开(公告)日：2019-10-17
申请号：US16381707
申请日：2019-04-11
Applicant: TuSimple
Inventor： Pengfei CHEN , Nan YU , Naiyan WANG , Xiaodi HOU
Abstract: Disclosed are devices, systems and methods for processing an image. In one aspect a method includes receiving an image from a sensor array including an x-y array of pixels, each pixel in the x-y array of pixels having a value selected from one of three primary colors, based on a corresponding x-y value in a mask pattern. The method may further include generating a preprocessed image by performing preprocessing on the image. The method may further include performing perception on the preprocessed image to determine one or more outlines of physical objects.
-
公开(公告)号：US20190301861A1
公开(公告)日：2019-10-03
申请号：US16290798
申请日：2019-03-01
Applicant: TuSimple
Inventor： Naiyan WANG , Yuanqin LU
Abstract: The present disclosure provides a method and an apparatus for binocular ranging, capable of achieving an improved accuracy of binocular ranging. The method includes: extracting features from a left image and a right image to obtain a left feature image and a right feature image; selecting a standard feature image and obtaining a cost volume of the standard feature image by applying a correlation calculation to the left feature image and the right feature image using a block matching algorithm; obtaining a confidence volume by normalizing computational costs of all disparity values in a disparity dimension for each pixel point in the cost volume; obtaining a confidence map by selecting a maximum value from confidence levels of all the disparity values in the disparity dimension for each pixel point in the confidence volume; obtaining a mask map by mapping each pixel point having a confidence level higher than a predetermined threshold in the confidence map to 1 and mapping each pixel point having a confidence level lower than or equal to the threshold in the confidence map to 0; obtaining a disparity map by calculating an argmax value for the confidence levels of all disparity values in the disparity dimension for each pixel point in the confidence volume; obtaining a target disparity map by multiplying the mask map with the disparity map; and estimating a distance based on the target disparity map.
-
-
-
-
-
-
Search Results
7
records
(took 0.018 seconds)
