Product

Use Case

Resource

Company

Partner

Contact us

Classification inspection solution

We accurately classify types and defects through an AI deep learning-based classification solution.

Neurocle's Solution Clients

Try accurately distinguishing similar defects using a classification inspection solution and automate the judgment criteria.

It distinguishes between different types based on the characteristics such as shape, color, and pattern of the subject, and determines them according to predefined classification criteria.

How to apply the inspection

-

See real examples

-

Type

Subtitle

Title

Type

Subtitle

Title

Type

Subtitle

Title

How to apply the inspection

-

See real examples

-

Type

Subtitle

Title

Type

Subtitle

Title

Type

Subtitle

Title

Main features

The Neurocle classification test solution provides a strong differentiator.

Precision classification of similar defects and types

Objects and defect types that are very similar in shape and characteristics, which are difficult to distinguish using existing inspections, are clearly classified by learning subtle differences.

Precision classification of similar defects and types

Objects and defect types that are very similar in shape and characteristics, which are difficult to distinguish using existing inspections, are clearly classified by learning subtle differences.

Multi-classification and multi-type classification

It simultaneously handles various product types and multiple defect types with a single classification model, performing consistent and efficient type classification even in complex inspection environments.

Multi-classification and multi-type classification

Stable classification even in the presence of defect data constraints.

Even in environments where it is difficult to obtain defect data due to high yield, we reliably implement high classification performance through normal data-centric learning and anomaly pattern recognition.

Stable classification even in the presence of defect data constraints.

Field Application Case

Application cases of external inspection solutions created with rich field experience

Red ginseng grade classification test

문제점

The image of ginseng captured using X-ray technology did not have significant color contrast, making the internal cavities indistinct. As a result, it was difficult for the human eye to accurately assess the grade of the ginseng, which led to issues regarding the trustworthiness of the product.

솔루션

We have solved the problem by introducing a vision inspection solution where a deep learning model recognizes the characteristics of red ginseng in X-ray images to determine its grade. We achieved an accuracy of 97% in grade determination, enhancing the ability to maintain the quality of red ginseng compared to before.

Red ginseng grade classification test

문제점

솔루션

Color-coated steel sheet surface texture inspection

문제점

In the inspection of distinguishing the texture differences of colored steel plate surfaces, it was difficult to make clear classifications when the differences were subtle. Additionally, since the inspection results relied on the skill level of the workers and subjective judgments, there was a lack of consistency, and there were limitations in ensuring quality reliability.

솔루션

We have quantitatively analyzed the fine roughness deviations, which were difficult to determine using traditional visual inspection and rule-based algorithms, through an automated deep learning model. As a result, we were able to reduce the possibility of non-compliance with customer specifications and lower the incidence of quality claims.