Verification of Optical Image Qualities

For COFs (Chip On Film) that are put into mobile devices and other electronic devices, both humans and machines are engaged in quality control using optical images. We have carried out image-based deep learning projects by utilizing COF images that are classified into optical images and applying various deep learning models in order to come up with satisfactory results. We own classifying modules for images used for deep learning and teaching systems.

In addition, we are getting many calls for applying deep learning into various fields, and capable of selecting parts that can effectively apply deep learning and deliver great results.

Deep Learning Application in Semiconductor Production Processes

Wafers that are used for semiconductors see its exponential quality declining if their waiting time in logistics processes get longer. It causes tremendous losses to semiconductor production companies, but there is no sure-fire solution. However, we have provided solutions based on deep learning to semiconductor companies with which we have worked. Differently from image based learning methods, we put our focus on input data processing and thereby, we have optimized logistics processes.

Estimation of Boiler Pipe Life Span

Boiler life span is determined by pipe life span. We worked with power generating companies to apply deep learning and to come up with estimated life span of pipes and to produce meaningful results. For quality inspection for boiler life span, we used pipe images for deep learning, and developed our own deep learning models.