Our Goal

Establish an automatic intelligent-driven inspection process

Dinnar has recently started to form a specialized team to research and develop various algorithm models from the academic community and adapt them to the industrial inspection field.
This initiative addresses the current shortcomings in the intelligence of traditional industrial vision inspection algorithms and the difficulty in feature extraction.
Since its formation, the team has completed dozens of actual projects, covering defect detection in various materials such as plastics, metals, and glass.

How to comprehend?

Top-Level Concept and Starting Point

Dinnar has created a production manufacturing optimization system based on data analysis and machine learning technologies. By employing technology, it addresses issues in the manufacturing process, enhances production efficiency, and is committed to providing reliable, efficient, safe, and maintainable technology to help manufacturers succeed in today's competitive market.

DN Framework

Application of technology + unique industrial data for rapid implementation of results

New generation machine learning platform and algorithms
  • Visual image + defect feature set

    These technologies enable accurate detection and identification of defects and errors during the production process, thereby improving product quality and efficiency.
  • Unique production data training

    Defect data, process data, production line specific data
  • Rapid optimization

    Enhancement of visual efficiency and model accuracy
General visual software Computing power Data processing center Application scenarios
  • Rapid implementation

    Technology applied in intelligent inspection devices can quickly achieve automated detection and analysis, improving production efficiency and product quality.
  • Cost reduction

    Technology can reduce labor investment and equipment costs, lower inspection costs, and enhance enterprise profitability.
  • Yield improvement

    Accurately identify product defects and anomalies, promptly detect and resolve issues during production, and improve product pass rate and production efficiency.

Leading industry algorithm capabilities

Defect sample generation model based on Generative Adversarial Networks
Effectively addresses issues of insufficient samples and sample imbalance Strong transfer capability
Generative fake sample set, but trained as a real sample set
Weakly supervised graph network, segmentation framework handles multiple weak annotations
Combined with strong supervision models, significantly improves performance Adapts to various weak annotations, significantly reducing costs
Bingfeng Zhang, Jimin Xiao, et al. "Affinity Attention Graph Neural Network forWeakly Supervised Semantic Segmentation." IEEE PAMI(2021)

Quality control

Intelligent quality inspection

With machine vision inspection, quickly scans product quality
High efficiency, low cost, high precision (up to micrometer level)
Potential to replace 35 million visual inspection workers globally

Production process

Process optimization

Establishes product health models through machine learning,
automatically matches the best process parameters
Improves product quality, optimizes effect in a closed loop

Intelligent production

Significantly improves production efficiency & quality, reduces costs

Decision optimization

Risk prediction

Shifts from post-event to pre-event prediction, significantly reducing the probability of risk events

Intelligent decision-making

Relies on massive industrial data and training
Significantly improves the accuracy & efficiency of production decisions

Why us?

Enhance efficiency and reduce costs, optimize product quality, enhance brand value

  • Enhance quality control

    Improves the level of product quality control, reduces the rate of defective products and customer complaint risks
  • Efficient production

    Achieves efficient production, reduces resource waste and rework
  • Support customization

    Customization capability and flexibility meet the needs of different industries and application scenarios
  • Compatibility

    Seamlessly integrates personalized solutions with existing systems
  • Sustainability

    Contributes to the sustainable development of the industry

Consultation Contact Us

  • Phone

    0512-66957689
  • Address

    No. 11 Tingxin Street, Industrial Park, Suzhou
  • Email

    info@dinnar.com