公开(公告)号：WO2023018369A2

公开(公告)日：2023-02-16

申请号：PCT/SG2022/050531

申请日：2022-07-26

Applicant: LEMON INC.

Inventor： TAY, Michael Leong Hou ,  MA, Wanchun ,  CHENG, Shuo ,  WANG, Chao ,  LUO, Linjie

IPC: G06V40/16 ,  G06V10/82 ,  G06N3/09 ,  G06T7/246 ,  G06F18/2193 ,  G06T13/40 ,  G06T13/80 ,  G06T2207/20084 ,  G06T2207/30201 ,  G06T7/251 ,  G06V10/242 ,  G06V40/171 ,  G06V40/176

Abstract: The present disclosure describes techniques for facial expression recognition. A first loss function may be determined based on a first set of feature vectors associated with a first set of images depicting facial expressions and a first set of labels indicative of the facial expressions. A second loss function may be determined based on a second set of feature vectors associated with a second set of images depicting asymmetric facial expressions and a second set of labels indicative of the asymmetric facial expressions. The first loss function and the second loss function may be used to determine a maximum loss function. The maximum loss function may be applied during training of a model. The trained model may be configured to predict at least one asymmetric facial expression in a subsequently received image.

公开(公告)号：WO2022002262A1

公开(公告)日：2022-01-06

申请号：PCT/CN2021/104308

申请日：2021-07-02

Applicant: 深圳思谋信息科技有限公司 ,  上海思谋科技有限公司

Inventor： 杨志成 ,  李睿宇

IPC: G06K9/32 ,  G06K9/6267 ,  G06V10/242 ,  G06V20/62 ,  G06V30/10

Abstract: 一种基于计算机视觉的字符序列识别方法、装置、计算机设备和存储介质。所述方法包括：获取携带有待识别字符序列的图像；基于预先构建的位置检测模型，获取图像中待识别字符序列所在的目标区域图像；将目标区域图像进行水平校正，得到水平的目标区域图像；基于预先构建的角度判断模型，获取水平的目标区域图像的字符序列摆立状态；若字符序列摆立状态为正立状态，则将水平的目标区域图像输入预先构建的内容识别模型，获取待识别字符序列对应的字符序列内容。

公开(公告)号：WO2022189482A2

公开(公告)日：2022-09-15

申请号：PCT/EP2022/055984

申请日：2022-03-09

Applicant: DETECTSYSTEM LAB A/S

Inventor： LOVMAND, Jan ,  TORP, Lars ,  HOLM, Rasmus Bækgård

IPC: G06V10/24 ,  G06Q40/00 ,  G06T7/00 ,  G06V10/32 ,  G06V10/36 ,  G06V10/56 ,  G06V10/74 ,  G06V20/30 ,  G06V10/86 ,  G06Q10/10 ,  G06Q40/08 ,  G06V10/242 ,  G06V10/761 ,  G06V2201/11

Abstract: The present invention provides a method for quantitatively rating the degree of similarity of images by using orientation, resizing of the digital image, pixilation of the digital image, creation of an image string suitable for generating a hit-score. In the generation of the image string color codes e.g. by the RGB system is used. The hit-score can be a percentage identity or homology between the image strings of the two digital images, said hit-score being the rating of the similarity between the two digital images. The invention also related to the use of the method in process of assessing insurance claims. Further the invention to the use of said method and for a computer program with instruction for carrying out the method.