| AOI |
AI verify host |
QMB |
2/5 Wait for customer send out training image
3/12 Meeting
-Pass rate need to improve
-AI model need to update
4/2 Kulanat: Schedule to update AI verify host
4/14 update RSAI version at F7 line O17
4/18 Upload D:\ai_image and ai_golden_image
4/24 Collect and emailing AI log and bkxml file. Scott reply setting was collect waiting for complete version after training finish.
5/8 Performance accepted, they will apply to all AOI.
Will close next week |
nooknick |
closed |
| AOI |
RS no image |
NA |
05/07 Send information to RD help to check abnormal.
05/08 Joyce emailing solved the problem follow manual first about Memory not enough
5/15 Waiting for customer inform stop the production line
5/19 Solved the problem follow SOP and Keep monitoring line K27 CS |
nooknick |
on going |
| AOI |
AI verify host demo |
Chicony |
6/5/2026 - Set up verify host pc and test run one machine. The AI verify host pc received data normally but the image is not showing. Already mailed to Taiwan wait to apply the solution
7/5/2026 - Apply the RS AI interface key to RS pc and change the module setting inside RS software from AI client to AI interface. the image is now display on the RSAI software. Only defect images can be displayed today, already sent an email to RD still waiting for solutions.
8/5/2026 - The path that RSAI pc read and the actual path is not the same, RD suggest to rename all RS pc to be different. If this solution does not work then have to modify the system file. Today have to connect both pre flow and post flow to the RSAI. Both Pre and Post reflow RS have been connected to the RSAI pc, the golden image can display normally and both golden and defect image can be save to the ai_image and ai_golden image folder.
11/5/2026 - Continue to collect the data from preflow and postflow AOI. Customer wants to set up AI at the other production line have to discuss futher(total 4 AOI machines 2 Preflow and 2 Postflow)
12/5/2026 - Monitor the data and check the AI report in SFC to compare the difference.
18/5/2026 - Get the AI report from SPC to make report for line 6. For line 5 have to gather more data. Send some of the training dataset to RD team to train model. AI accuracy for line 5 Pre and Post flow is at 76%. Wait for customer decision to turn on AI auto confirm. |
phone |
on going |
