Application scenario: Applied to visual inspection of process. The product have been applied in 10+ projects in display panel industry
Product function
– Object detection、Image classification、Image segmentation
– Model management: train、inference
– Data management
– Senior judgement
– Monitoring and early warning
– High availability
Product features
– Reduce quality inspection cost
– Precipitate knowledge and skills
– Improve detection capability dynamically
– Promote comprehensive process improvement
Advantage effect
|
Customer oriented product design, which can reduce the quality inspection cost by 85% and promote the comprehensive improvement of process |
|
The training process is visualized, the model results are clearly visible, and the model effect is clear at a glance |
|
Manual judgment, to complement the automatic judgment |
|
Rich data analysis functions, with front layer traceability, map analysis, similarity analysis, decision analysis and other functions |
Cases
※A international leading pan-semiconductor enterprise case
|
• Business challenges
1. Many kinds of defects (100+) and difficult to sort 2. Manual judgment with low accuracy and efficiency 3. Lack of AI talent
4. Traditional thinking 5. Lack of methods to find out high-value and applicable scenarios
|
• Solution
|
• Effect of solution
1、Automatically judges the map, the personnel replacement rate is 80%, and the accuracy rate is improved by 20% 2、Reduce quality inspection cost by 85% 3、Improve product quality, reduce non-quality cost (scrap, etc.) 4、Provide root cause analysis of product defects 5、5 + people have learned to use and iterate the existing machine learning model 6、Established the flow system of defect detection
|
Advantage effect
|
Customer oriented product design, which can reduce the quality inspection cost by 85% and promote the comprehensive improvement of process |
|
The training process is visualized, the model results are clearly visible, and the model effect is clear at a glance |
|
Manual judgment, to complement the automatic judgment |
|
Rich data analysis functions, with front layer traceability, map analysis, similarity analysis, decision analysis and other functions |
Cases
※A international leading pan-semiconductor enterprise case
|
• Business challenges
1. Many kinds of defects (100+) and difficult to sort 2. Manual judgment with low accuracy and efficiency 3. Lack of AI talent
4. Traditional thinking 5. Lack of methods to find out high-value and applicable scenarios
|
• Solution
|
• Effect of solution
1、Automatically judges the map, the personnel replacement rate is 80%, and the accuracy rate is improved by 20% 2、Reduce quality inspection cost by 85% 3、Improve product quality, reduce non-quality cost (scrap, etc.) 4、Provide root cause analysis of product defects 5、5 + people have learned to use and iterate the existing machine learning model 6、Established the flow system of defect detection
|