DAP Vision
This is a deep learning-based vision inspection service for determining defects in manufacturing lines.
LG CNS's DAP Vision
DAP Vision is a vision inspection service that offers efficient, convenient learning service through a high-performance GPU-based platform and learning process automation function and provides the factory site with a judgment service system that can quickly implement deep learning judgment models to enable continuous performance management through judgment model deployment, implementation, and relearning.
As more images are available on site, the accuracy of deep learning vision inspections is improved accordingly, and DAP has implemented a process that can provide a seamless process through image acquisition-learning-model generation-deployment.
DAP Vision(Data Analytics & AI Platform Vision)
Service Features
-
01Optimal model selection learning system
The DAP Vision service has built a learning system in the cloud based on high-performance GPUs. It automatically collects factory-judged images and result data and automatically distributes and manages the image sets according to purpose. based on collected data, various frameworks (Tensorflow, Caffe, and AutoML) and deep learning networks can be selected to train the model. The easy learning function makes it easy for anyone to handle a multi-level learning process with a single request. Snapshot models created after training can also be compared using process control indicators such as untested, overtested, and contribution rates to select optimal models for production lines based on reference values.
-
02Excellent management function and decision system
DAP Vision's local system is divided into management server and decision servers. This structure can manage models distributed to multiple decision servers through a single management server. The online distribution function enables users to mount and manage models determined from the learning system in a large number of production lines. Judgment results are provided in real time; you can check the judgment statistics and judgment performance as well as summary information on the dashboard through the preset batch statistics.
-
03Vision inspection environment linked with Google AutoML
The DAP Vision service is linked with Google's AutoML. AutoML makes it easier even for non-experts to conduct deep learning and create models with fewer images. The Google EdgeTPU (beta service) optimized for the neural network process can be applied to the judgment system, and the judgment server can be implemented at a lower cost.
Core Functions
Our deep learning-based AI vision inspection service improves productivity with a vision inspection platform that automates the inspection process at all stages of production, from raw materials entrance into factories to final assembly securing inspection accuracy and improving productivity.
Solution Architecture
References
See the customers using LG CNS's
DAP Vision.
LG ChemMicro-cell quality inspection
LG ChemSynthetic rubber quality inspection
LG ChemGlass substrate quality inspection
LG ChemOptical material quality inspection
LG DisplayPanel quality inspection
Platform
-
DAP
A platform that integrates big data analysis and AI service development environment and enables operation at customer sites by using trained models.
Solution
-
DAP Talk
Anyone can produce solutions quickly and easily using conversation services that supports both voice and text.
-
DAP MLDL
This is an enterprise data platform that supports the entire life cycle from data lake to production.
-
SBP (Smart Big Data Platform)
This is a Hadoop-based Big Data platform optimized for domestic corporate services with enhanced reliability and accessibility through innovative enterprise-targeted features.
-
Smart SMA (Social Media Analytics)
This analysis platform utilizes social big data through consumer analysis such as status data of customer markets, companies, and products and services.
-
Wish to connect directly with the personnel?
Contact Information
- DAP Vision
- 02-2099-3111
- shku@lgcns.com
“*” indicates a required input field.
Close Popup-
*Name
-
*Email
@
-
*Company (department)*Job PositionDepartment name
-
Intent of Application
-
I Agree
Click the link below and read the content carefully. By checking I Agree, you are deemed to have read and agreed to all of the following.