Product

Use Case

Resource

Company

Partner

Contact us

Visual Inspection Solution

An AI deep learning-based inspection solution that
quickly and accurately detects defects in product appearance.

Neurocle's Solution Clients

Improve process productivity and product quality with exterior inspection solutions.

We accurately identify defects or imperfections in the visible exterior aspects of products or components, such as the surface, shape, and color.

Detectable information

Scratch

Crack

foreign substance

pollution

Transformation

Label defect

Scratch

Transformation

Crack

Label defect

foreign substance

pollution

Scratch

foreign substance

Transformation

Crack

pollution

Label defect

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 powerful differentiator provided by the Neurocle appearance inspection solution.

Accurately detect even the slightest differences.

It precisely identifies micro defects at the micron level, accurately determining defects that are difficult to distinguish with the naked eye.

Accurately detect even the slightest differences.

It precisely identifies micro defects at the micron level, accurately determining defects that are difficult to distinguish with the naked eye.

Simultaneously detecting various types of defects

Regardless of the form, number, or type of defects, multiple types of defects can be detected simultaneously in a single inspection.

Simultaneously detecting various types of defects

Regardless of the form, number, or type of defects, multiple types of defects can be detected simultaneously in a single inspection.

Building inspection models with a small amount of defect data

Generate virtual data similar to actual defects to implement a stable appearance inspection model even in a limited defect data environment.

Building inspection models with a small amount of defect data

Generate virtual data similar to actual defects to implement a stable appearance inspection model even in a limited defect data environment.

Field Application Case

Application cases of external inspection solutions created with rich field experience

Cylinder head defect inspection

문제점

We were conducting inspections to detect defective holes in the exterior of the car cylinder. The visual inspections being carried out previously were prone to human error, and the rule-based inspection method resulted in a large number of over-checks.

솔루션

We have equipped the existing inspection equipment with a deep learning model to detect defects on the exterior of engines. We have increased the production of cylinder heads from 300 to 400 units per day (a 33% increase in production rate), and the defect detection rate has risen to 96%.

Cylinder head defect inspection

문제점

솔루션

Semiconductor wafer surface defect inspection

문제점

It was necessary to conduct an inspection to find defective areas on the surface of semiconductor wafers to determine whether they were good or bad. A deep learning model was implemented for inspecting semiconductor photoresists, but there was a problem where defects disappeared in high-resolution images.

솔루션

By using a model that inspects high-resolution images divided into patches, we have been able to accurately detect defects as small as 5μm without resizing the high-resolution images.

Semiconductor wafer surface defect inspection

문제점

솔루션

By using a model that inspects high-resolution images divided into patches, we have been able to accurately detect defects as small as 5μm without resizing the high-resolution images.