December 28, 2020
Adacotech begins offering a closed beta version of “AdaInspector Cloud” a SaaS system that automates the inspection and testing of manufacturing products.
On December 28, 2020, Adacotech Co., Ltd.(Head office: Chiyoda-ku, Tokyo, CEO: Ryota Kawamura, hereinafter referred to as “Adacotech”) , launched the closed beta version of AdaInspector Cloud(* 1), a SaaS system that automates inspection and inspection at manufacturing sites without the need for programming.
Adacotech is a technology company that provides software that enables more efficient anomaly detection than ever before, based on image analysis technology using the HLAC(* 2) feature extraction method developed by the National Institute of Advanced Industrial Science and Technology.This technology enables highly accurate analysis with a small amount of data, and because the model “detects deviations from normal” rather than detecting what has been learned as abnormal, it can detect almost 100% of abnormalities, including unprecedented cases.
In addition, since it is a simple method that can be calculated using only product-and-sum operations, the burden of calculation processing is small. And another advantage is that the detailed inspection process can be realized and operated on a general-purpose PC like a notebook computer.To date, we have already installed image analysis solutions for inspection and inspection on the actual production lines of several companies, including major automotive parts manufacturers, and we are promoting the practical use of our technology mainly in the automotive industry.
Adacotech has been building and providing on-premise systems customized to meet the needs of each company to automate inspection and testing at manufacturing sites.The quality requirements of the manufacturing industry are very strict, and at the manufacturing site, advanced inspection and testing is carried out under strict inspection standards to eliminate the oversight of defective products.Therefore, the automation of inspection and inspection using AI-based image analysis requires the knowledge and skills of skilled engineering professionals.
On the other hand, there is a growing shortage of engineers in Japan, and in order to meet the growing need for automation of inspection and testing in the manufacturing industry, it is necessary to provide cloud services that allow users to build their own systems.In addition, in order to realize automation of inspection and testing at more manufacturing sites, it is necessary to provide systems with low implementation costs and low operational management burdens.Adacotech is developing the SaaS system “AdaInspector Cloud” in order to provide a service that solves these issues and contributes to the improvement of productivity in the manufacturing field. Prior to the launch of the official version, a closed beta version was launched on December 28, 2020.
AdaInspector Cloud” is a SaaS system that provides integrated construction and implementation of automated inspection and testing systems using image analysis.It consists of a service that automatically learns and builds a system in the cloud, and an inspection application that runs on the manufacturing floor, and is scheduled to launch around the summer of 2021.Normally, when automating inspection and testing in the manufacturing industry through image analysis using AI, it is necessary to optimize the pre-processing of the product images to be inspected, optimize processing of the learning method by combining multiple patterns, and achieve performance and processing speed. Developers are required to have programming skills.In addition, in order to maximize the inspection accuracy of the constructed system, it is essential to adjust the parameters(*3) and evaluate the accuracy, and these tasks need to be repeated in a trial-and-error manner until they are put into practical use, requiring the skilled skills and knowledge of engineers.AdaInspector Cloud” is a system that enables the construction of highly difficult inspection and testing automation systems and their execution on the manufacturing floor without programming and with simple and intuitive operations.Even if you are not specialized in engineering in the area of AI or image analysis, we are developing an automated system for inspection and inspection in-house, with the aim of making it possible to operate in the field.
The closed beta version of “AdaInspector Cloud,” which is now available ahead of the official version, mainly provides functions to generate trained models(*4) using the company’s own product images prepared by the user, and functions to evaluate the accuracy of the generated trained models.By simply uploading about 100 images of the product to be inspected or inspected, you can complete the data set (*5) to be used for training and testing the model.
It also provides an “automatic learning function” that generates a model to automate inspection and testing by executing learning with only a few steps, and a “model testing function” that quantitatively shows the inspection accuracy of the trained model created by the “automatic learning function.Adacotech will use the feedback from the beta version users as a reference to improve the product and increase its convenience in preparation for the launch of the official version in 2021. Adacotech aims to develop a business that supports the evolution and innovation of the manufacturing industry through the spread of automated inspection and testing.
* 1″AdaInspector Cloud” closed beta version is available on a limited basis and is not open to the public at this time.
* 2HLAC feature extraction method:A general-purpose and fast feature extraction method with excellent recognition accuracy used for image analysis and recognition.In contrast to the Deep Learning technology, which requires complex processing when calculating the shape and size of the inspection target, it can be calculated by simply performing a sum-of-product operation on the values representing the color shade and brightness of each pixel, which can be instantly calculated on a commercial PC.It is also capable of recognizing the same object even if its position has changed, and when there are two objects, the sum of the features of each object becomes the overall feature.In other words, there is no need to find the boundaries of regions that represent the same thing in an image (segmentation free), and the features can be recognized individually even when multiple anomalies occur in an image.It is a feature extraction method with desirable properties for image recognition.
* 3A parameter translated as “variable,” is a value that is adjusted to the algorithm in order to improve the performance and accuracy of learning when performing machine learning or deep learning.
* 4A learned model is a specific method of calculation created through “learning” by using an algorithm to analyze data in a specific field to find regularities and relationships among them.
* 5A data set is an aggregation of data that is processed by a program used to train and test models in the process of building AI.