Started providing GA version of AI cloud service “ADFI” that can create visual inspection AI without coding.
AI Robotics Ltd., which provides AI cloud services, will start offering the GA version of the cloud service “ADFI” that can create visual inspection AI with no code from August 31, 2022.
Currently, automation of visual inspection is attracting attention as DX in the manufacturing industry, but the high cost of developing visual inspection AI is a barrier to introduction. In general, visual inspection AI is developed over several months by a system engineer with AI expertise. In addition, it is necessary to collect new learning data for each inspection target and develop a new AI, making it difficult to reuse already developed AI. Due to these factors, the introduction of visual inspection AI costs a lot of money, so in most small and medium-sized factories, the introduction has been abandoned.
ADFI makes it possible to create visual inspection AI easily and free of charge, even without specialized knowledge of AI, which significantly reduces development costs. By using ADFI, it is possible to introduce visual inspection AI even in cases where introduction was abandoned due to high costs.
Features of ADFI
(1) Image anomaly detection AI can be created and performance verified on the same day simply by operating on the screen.
After creating an account, you can immediately create an image anomaly detection AI and perform its performance verification on the screen of a web browser. All you need to do is upload training images and test images and press the training execution button. Troublesome AI parameter adjustment is unnecessary. Also no AI expertise or expertise required.
(2) Automatically create the world’s most advanced high-performance AI.
ADFI is equipped with the world’s most advanced AI algorithm, and can automatically create extremely high-performance image anomaly detection AI. In a performance evaluation experiment using image anomaly detection benchmarks, we achieved accuracy equal to or better than that of the three world-leading methods as of 2021.
(3) Original technology requires only a small number of images for AI learning.
ADFI can create an image anomaly detection AI with less than 1/10 the amount of image data required for learning general deep learning (minimum 10 normal images only).
(4) No need for expensive hardware.
ADFI trains and executes AI on the cloud, so users do not need expensive computers such as GPU-equipped servers.
Free Trial Tutorial
You can download the tutorial and try ADFI for free!
Tutorial URL: https://adfi.jp/wp-content/uploads/ADFI_Free_Trial_Tutorial.pdf