Use Cases and Benefits of Image Judgment AI for Manufacturing Industry

1. What is Image Judgment AI?

Image judgment AI is a technology where AI analyzes images captured by cameras or sensors to automatically detect defects or abnormalities such as scratches, dirt, and chips. Traditionally, visual inspections by humans and rule-based image processing were mainstream, but with the advent of AI (especially deep learning), more accurate and flexible judgments have become possible.

In deep learning, by training on a large amount of image data, it becomes possible to recognize subtle defects and complex patterns that are difficult for humans to notice. Therefore, it is highly effective against traditional issues such as "ambiguous inspection criteria" and "variability among inspectors," and a major feature is its ability to adapt to changes in the inspection environment and differences between products.

In particular, in the manufacturing industry, adoption is progressing in quality control areas such as visual inspection, contributing to solving labor shortages and standardizing inspection accuracy.

Reference Blogs:
What is Image Recognition? Mechanisms of Image Recognition and Examples of AI Applications
What is Deep Learning? Introducing Differences from Traditional Machine Learning and Key Points for AI Business Implementation.
Improving Work Efficiency with AI. Four Machine Learning Implementation Cases Where 80% Experience Benefits.

2. Specific Implementation Examples in Manufacturing

The utilization patterns of image judgment AI in the manufacturing industry are truly diverse. It not only improves the efficiency of inspection processes but also contributes to line balancing across the entire production line, stabilizing quality, and preventing foreign object contamination. Here, we will introduce case studies of image judgment implementation.

Case 1: Aikawa Press Industry – Introduction of Automated Visual Inspection AI Software

At Aikawa Press Industry Co., Ltd., an automated visual inspection system utilizing image analysis technology was introduced for terminal plates used in electrical circuits of automotive electrical components called "flat bus bars," which had been difficult to detect defects by system and were inspected visually by skilled workers until now.

Adacotec provides automated visual inspection AI software to Aikawa Press Industry's automotive parts production line

Case 2: Kewpie – Deployment of AI-Powered Raw Material Inspection Equipment Across the Group

Kewpie introduced an in-house developed raw material inspection device utilizing AI for inspecting cut vegetables used as ingredients in prepared foods, automating the inspection that was previously done visually. The task of removing foreign objects mixed in the raw materials by human inspection placed a significant physical burden on workers, and automating this process has improved work efficiency.

Deployment of AI-Powered Raw Material Inspection Equipment Across the Group

Case 3: Ito Metal Industry – Automation of Defective Bolt Product Judgment with AI System

Automotive parts manufacturer Ito Metal Industry has introduced a system using AI technology for the visual inspection of bolt unions. The company had been using automated visual inspection machines for 15 years, but due to issues such as false positives where good products were detected as defective and the inability to accommodate new defect judgment criteria, they decided to implement a new system aimed at improving judgment accuracy.

Ito Metal Industry, an automotive parts manufacturer, automates defect judgment of bolt products with an AI system

3. Key Points and Precautions When Implementing

By introducing AI, significant reductions in inspection man-hours and operational costs, as well as stabilization of quality, can be expected. However, to achieve sufficient effects, prior planning and designing the implementation process are essential. Below, we explain the key points and considerations to keep in mind when introducing image judgment AI in manufacturing sites.

1. Collecting and Preprocessing Inspection Data is Key
The accuracy of AI depends on the training data. It is important to prepare a sufficient number and quality of image data at the initial stage. Additionally, by clearly defining annotation rules and preparing training data that covers various variations of the target product images, the model’s accuracy can be improved.

2. Small Start with ROI in Mind
Rather than implementing across the entire line at once, start with a small-scale pilot operation to verify the ROI (Return on Investment). By identifying issues and making improvements in the initial phase and then gradually expanding, you can maximize the benefits of implementation while minimizing risks.

3. Data Protection and Security
Image data may contain product information and intellectual property. When using the cloud, be sure to enforce encryption and access controls, and clearly define data management rules both inside and outside the company to pay maximum attention to data protection and security.

4. Model Maintenance and Continuous Learning
When manufacturing lines or product specifications change, not only can the AI model's recognition accuracy decline, but it is often the case that the model tailored to the specific product cannot be reused as is. It is necessary to establish a system for regular model updates and additional learning to ensure stable long-term operation.

5. Involving Internal Stakeholders
By collaborating not only with the on-site team but also with the quality control and production management departments, and unifying AI operation rules and decision criteria, the adoption rate can be improved. Understanding and cooperation from stakeholders are the keys to successful operation.

4. Summary

Image judgment AI in manufacturing is a powerful tool that simultaneously achieves improved inspection accuracy, enhanced work efficiency, and cost reduction.

Furthermore, the accumulation of inspection data through AI implementation can also be utilized for visualizing quality control and making decisions for process improvements. This goes beyond mere inspection automation, serving as a foundation for optimizing the entire manufacturing process and future smart factory initiatives.

The introduction of AI is expected to be utilized not only in visual inspection but also in broader areas such as assembly processes, transportation processes, and preventive maintenance. When considering AI implementation, a strategy of starting small to verify ROI and gradually expanding the scope of application is effective. As a measure to enhance the competitiveness of manufacturing sites while balancing quality and productivity, image judgment AI will play an increasingly important role in the future.

5. Human Science Teacher Data Creation and LLM RAG Data Structuring Outsourcing Service

Extensive Track Record of Creating 48 Million Pieces of Training Data
At Human Science, we participate in AI model development projects across a wide range of industries, starting with natural language processing and extending to medical support, automotive, IT, manufacturing, and construction. Through direct business with many companies, including GAFAM, we have provided over 48 million pieces of high-quality training data. From small-scale projects with just a few members to large, long-term projects with a team of 150 annotators, we handle various types of training data creation, data labeling, and data structuring regardless of industry.

Resource Management Without Using Crowdsourcing
At Human Science, we do not utilize crowdsourcing; instead, we advance projects with personnel directly contracted by our company. We carefully assess each member's practical experience and their evaluations from previous projects to form a team that can deliver maximum performance.

Not Only Creating Training Data but Also Supporting Generative AI LLM Dataset Creation and Structuring
We support not only labeling for data organization and the creation of training data for identification-based AI but also the structuring of document data for generative AI and LLM RAG construction. Since our founding, we have been engaged in manual production as a primary business and service, and we provide optimal solutions leveraging our unique expertise and deep knowledge of various document structures.

Equipped with an In-House Security Room
At Human Science, we have a security room within our Shinjuku office that meets ISMS standards. Therefore, we can guarantee security even for projects handling highly confidential data. We consider ensuring confidentiality to be extremely important in every project. Even for remote projects, our information security management system has received high praise from clients, as we not only implement hardware measures but also continuously provide security training to our personnel.

In-House Support
Our company also provides personnel dispatch services for annotation-experienced staff and project managers who match the customer's tasks and situation. It is also possible to organize a team under the customer's on-site supervision. Additionally, we support the training of your workers and project managers, selection of tools tailored to your situation, automation, work methods, and the construction of optimal processes to improve quality and productivity, assisting with any challenges you face related to annotation and data labeling.