S&V Measure AI Solution

Doberman Explorer 

Converts a component’s sound measurement data extracted from 

test equipment to images

Doberman Trainer 

Analyzes input image data and builds up the right measurement algorithm 

for a component

Doberman Detective 

Determines whether products are defective or not, using its unique measurement algorithm

Quality inspection of mass-production products
Quality inspection of mass-production products
Precision Special Measuring Device
Precision Special Measuring Device
Telemetry and Surveillance
Telemetry and Surveillance
Quality inspection of noise vibration of products shipped from production line, manufacture/supply of products
Quality inspection of mass-production products
Automation of special noise measurement solutions such as aviation/defense/precision equipment, laboratory Q/A laboratory measurements, etc.
Precision Special Measurement Equipment
NoiseVibration Telemetry and Diagnostics Solutions, Network Data Measurement Solutions
Telemetry and Surveillance

 Doberman®is a sound measurement automation solution based on deep learning, which dramatically decreases human interference and an initial preparation process.According to the unique feature of a component to be measured, three programs are provided to decide whether it is defective or not. 


Doberman®is a sound measurement automation solution based on deep learning, which dramatically decreases human interference and an initial preparation process.
According to the unique feature of a component to be measured, three programs are provided to decide whether it is defective or not.

Doberman® Explorer

Converts a component’s sound measurement data extracted from test equipment to images

Doberman® Explorer


Converts a component’s sound measurement data extracted from test equipment to images


Extracts converting conditions that are proper to the unique trait of a component
Extracts converting conditions that are proper to the unique trait of a component
Converts a large quantity of sound data to images according to conditions
Converts a large quantity of sound data to images according to conditions

Doberman® Trainer


Analyzes input image data and builds up the right measurement algorithm for a component 


Inputs a large quantity of data to the training algorithm
Inputs a large quantity of data to the training algorithm
Calculates the degree of defect with the specialized algorithm
Calculates the degree of defect with the specialized algorithm

Doberman® Detective


Determines whether products are defective

 or not, using its unique measurement algorithm 



Calculates each component’s degree of defect according to the pre-set criteria by the Trainer, and decides whether it is defective 

Doberman® Trainer

Analyzes input image data and builds up the right measurement algorithm for a component 

Doberman® Detective

Determines whether products are defective or not, using its unique measurement algorithm 


Production line check function 

Doberman always provides a certain standard of judgement, so it is easy to identify problems in the production line. 


The most frequent value of an abnormal score deviates from the initial reference line. 

 → Need to review production line issues 


How Doberman®Works

The program learns the criteria by itself

Doberman uses the anomaly detection method based on deep learning to calculate “abnormal points” with a majority of data that have similarity. Every component is assigned with the “abnormal point” to make it easy to determine whether the component is defective or not.

Effects of Doberman®

Credible measurement is available with less time and cost

01 Credibility Enhancement 

  • It is able to detect unknown defects
  • It offers discrimination capacity at a constant level 

02 Cost Reduction 

  • Drastic reduction in personnel expenses and maintenance cost
  • Able to be applied to various models 

03 Time Reduction 

  • It minimizes the period of sorting out defective goods and collecting them

  • It automatically analyzes the criteria



From Production Management to Brand Value Management


Detecting defects in advance in the production process of home appliances and automobiles dramatically reduces the cost of managing customer claims and replacing complete products. It is also helpful for building up a credible brand image.* The case of the automobile company A



About 20% of the whole customer claims are on noise 


(After the 1st repair) The re-repair rate is about 46% 


About 32% of the products are sent to the central A/S center 

How Doberman®Works

The program learns the criteria by itself

Doberman uses the anomaly detection method based on deep learning to calculate “abnormal points” with a majority of data that have similarity. Every component is assigned with the “abnormal point” to make it easy to determine whether the component is defective or not.

Effects of Doberman®

Credible measurement is available with less time and cost

01 Credibility Enhancement
02 Cost Reduction
03 Time Reduction
  • It is able to detect unknown defects
  • It offers discrimination capacity at a constant level
  • Drastic reduction in personnel expenses and maintenance cost
  • Able to be applied to various models

  • It minimizes the period of sorting out defective goods and collecting them
  • It automatically analyzes the criteria



From Production Management to Brand Value Management


Detecting defects in advance in the production process of home appliances and automobiles dramatically reduces the cost of managing customer claims and replacing complete products. It is also helpful for building up a credible brand image.

* The case of the automobile company A



About 20% of the whole customer claims are on noise 
(After the 1st repair) The re-repair rate is about 46% 
About 32% of the products are sent to the central A/S center