公开(公告)号：US11195045B2

公开(公告)日：2021-12-07

申请号：US16848565

申请日：2020-04-14

Applicant: GETAC TECHNOLOGY CORPORATION

Inventor： Kun-Yu Tsai ,  Po-Yu Yang

IPC: G06T7/586 ,  G06T7/00 ,  G06N3/08 ,  G06T7/40 ,  G06T7/11 ,  G06T7/45 ,  G06N3/063 ,  G06K9/46 ,  G06K9/00 ,  G06K9/20 ,  G06K9/32 ,  G06K9/62 ,  G01N21/3581 ,  G01N21/88 ,  G06N3/04 ,  G01J3/28 ,  G01N21/956 ,  G06F17/16

Abstract: A method for regulating a position of an object includes detecting a plurality of first alignment structures of the object under rotation of the object, wherein a plurality of second alignment structures of the object sequentially face a photosensitive element during the rotation of the object, and when the plurality of first alignment structures have reached a first predetermined state, stopping the rotation of the object and performing an image capturing procedure of the object. The image capturing procedure includes: capturing a test image of the object, wherein the test image includes an image block presenting the second alignment structure currently facing the photosensitive element; detecting the position of the image block in the test image; when the image block is located in the middle of the test image, capturing a detection image of the object.

公开(公告)号：US11650164B2

公开(公告)日：2023-05-16

申请号：US16848576

申请日：2020-04-14

Applicant: GETAC TECHNOLOGY CORPORATION

Inventor： Kun-Yu Tsai ,  Po-Yu Yang

IPC: G06T7/00 ,  G06T7/40 ,  G06V20/64 ,  G01N21/88 ,  G06T7/586 ,  G06N3/08 ,  G01N21/3581 ,  G06N3/04 ,  G06T7/11 ,  G01J3/28 ,  G01N21/956 ,  G06T7/45 ,  G06F17/16 ,  G06N3/063 ,  G01N21/01 ,  G01N21/952 ,  G06V10/22 ,  G06F18/214 ,  G06N3/047 ,  G06V10/145

CPC classification number: G01N21/8806 ,  G01J3/2823 ,  G01N21/01 ,  G01N21/3581 ,  G01N21/8851 ,  G01N21/952 ,  G01N21/956 ,  G06F17/16 ,  G06F18/2148 ,  G06N3/04 ,  G06N3/047 ,  G06N3/063 ,  G06N3/08 ,  G06T7/0004 ,  G06T7/11 ,  G06T7/40 ,  G06T7/45 ,  G06T7/586 ,  G06T7/97 ,  G06V10/145 ,  G06V10/22 ,  G06V20/64 ,  G06V20/647 ,  G01N2021/8887 ,  G06T2207/10152 ,  G06T2207/20081 ,  G06T2207/20084

Abstract: An artificial neural network-based method for selecting a surface type of an object includes receiving at least one object image, performing surface type identification on each of the at least one object image by using a first predictive model to categorize the object image to one of a first normal group and a first abnormal group, and performing surface type identification on each output image in the first normal group by using a second predictive model to categorize the output image to one of a second normal group and a second abnormal group.