With SCAN-X we offer our customers a state of the art computer vision solution

Over the past few years, lexoro has specialised in computer vision, among other things. Since then, as an IT solution provider with SCAN-X, we have been taking on software projects in individual applications.

Below we outline current information on the field of computer vision and our technology SCAN-X. The download will take you to our latest white paper on the subject of Origin & Development of Computer Vision.

Origin & development of computer vision

Computer vision is a research area of artificial intelligence which deals with the question of how machines can see and understand digital images and videos. With the aim of approximating or exceeding the hit rate compared to the human level, this information should be made usable for further processes.

Research in this area of artificial intelligence began 60 years ago. At that time, two scientists – David Hubel and Torsten Wiesel – were able to prove via initial tests that visual pro- cessing always begins with simple structures, such as edges. In the 1960s, a project was started at MIT with a small group of students to develop a platform which would implement automatic level segmentation. Although the project failed, it is commonly referred to as the birth of computer vision. Supported by innovations in technology and performance, further improvements and developments in technology were carried out in the decades which followed.

However, the first real breakthrough came in 2012, when a team from the University of Toronto developed a new model – the so-called AlexNet. As a result, the error rate of image classification could be reduced from 26% to 16.4%.

Fig. 1: Statista 2020: error rates in image recognition from 2010 (28%) to 2017 (2.3%) compared with human ability (5%)

The Statista graphic above shows the progress which has been made since the “big breakthrough”. Computer vision was able to significantly exceed human capabilities and further potential for improvement can no longer be ruled out. Because of this, computer vision is extremely relevant in the company-wide AI environment, on the one hand due to its enormous performance, and on the other hand due to its individual options for application. The importance of com- puter vision is underpinned by a study published by “Markets and Markets” in their Market Research Report. The global computer vision market is forecast to grow from $11.9 billion to $17.4 billion between 2018 and 2023.

Features & status quo

Due to the ongoing improvements in technologies and possibilities, it isn’t easy to define a status quo. One reason for this is constant innovation with new use cases.
A suitable example is demonstrated by a breakthrough in computer vision which took place in the United States a few months ago: Researchers were able to recognise and determine 3D objects for the first time. The groundbreaking aspect here was training the model with deep learning. The system for recognising the object was “fed” exclusively with CAD data in order to train neural networks. Until then, there had to be a physical element which had to be photographed about 50 times to collect the necessary data.

One advantage of this procedure is the increased efficiency, since a significantly larger number of objects can be integrated into the system in a much shorter time.

Since the computer vision breakthrough in 2012, these kinds of innovative research results are no longer uncommon. WIPO (World Intellectual Property Organisation) has collected and published appropriate figures on this topic in its study on technology trends. Accordingly, there has been a “patent boom” in the field of artificial intelligence since 2012. Almost 50 percent of all patents related to AI mention computer vision, with an average annual growth rate of around 25 percent. It is therefore not surprising that computer vision comes first among functional applications.

Building on the two most relevant innovation factors (improved connectivity and better computer perfor-mance), various “building blocks” have been researched and developed in recent years which are used individually or in combination to solve complex problems and questions.

Fig. 2: Overview of the six features of computer vision

With SCAN-X technology, lexoro currently defines six “features”:

 

  • Object segmentation: automatic segmentation of the object from the background for digital processing operations

  • Object counting/triggering: Counting of typal objects

  • Automated fitting: automatic adjustment of different objects for fitting

  • Object & motion detection: automatic object and motion detection by scanning (e.g. with a smartphone)

  • Object Measurement: measurement of the object according too various attributes (length, depth, height, etc.)

  • 3D wireframing/digital cloning: generation of a point cloud for further use of the digital clone in various application environments

Artificial intelligence and its applications will meet the predicted growth potential in the years to come! This is supported by countless studies and forecasts from a wide range of industries and branches. The cause of this huge increase in sales has several factors. The overall catalysts are faster and faster computers which can process larger and larger amounts of data. As a result, more and more complex and costly tasks can be automated and optimised. Make use of the potential savings – made possible by artificial intelligence!

The key phrase “potential savings” is a broad but often-used term which should not be used as a basis for making decisions.

lexoro describes three main advantages which make it possible to objectively measure whether and to what extent artificial intelligence can optimise a company:

 

  • Cost and time savings
  • Precision & quality assurance
  • Independent learning & automation

The main reasons why companies are still reluctant to use artificial intelligence despite the clear advantages can be found in a study carried out by Deloitte in 2019.
They interviewed AI experts worldwide, including 100 decision-makers from German companies. The biggest risks and challenges are the lack of trust in artificial intelligence and its advantages, as well as data protection. Furthermore, according to the study, inadequate data quality and the inability to identify use cases, coupled with a lack of skills and specialists in the company, are the greatest challenges. Businesses need to do away with these concerns or take precautions to prevent cyber risks.

This is also confirmed by the results of a survey by adesso, in which 329 decision-makers were interviewed. The result: almost 50% see the use of new technologies such as AI as a key challenge for the next three years. This percentage makes it clear that a large number of German companies will implement AI applications in the coming years. It’s important to develop concrete plans and strategies at an early stage. Otherwise it is questionable whether competitiveness can be guaranteed in the long term.

Use computer vision for competitive advantages

Countless studies by various well-known institutions show that the upheaval and trend towards new technologies are coming or have already been implemented in many areas. The advantages are obvious, as are the risks. It is therefore important to use the positive aspects for competitive advantages and to counter the risks.

A considerable part of this is in computer vision. With almost 50% participation in AI patent applications, computer vision is already a crucial influencing factor in the AI world of the future. Driven by innovation, research and technological progress, groundbreaking achievements have been made in this field over the past decade. In the meantime, human capabilities are far exceeded by the possibilities of vision systems. As a result, increasingly complex projects with more difficult challenges can be implemented and solved automatically.

The most important thing here is not to implement artificial intelligence in the company or implement projects on this topic at all costs. On the contrary, market share and profit- ability should still be the main concerns.

This means that for lexoro, the return on investment (ROI) has a very high priority and importance. In combination with the afore-mentioned advantages – if implemented correctly – you can achieve enormous competitive advantages in order to generate sustainable added value for your company.

The decisive criteria: the correct implementation of the right use cases! The development of know-how and expertise is a costly and lengthy process, which in many cases doesn’t lead to the desired result.

Please make an appointment with us.

Interested?

Then contact us directly in person:

Fill out my online form.

Or call us.

Sebastian van der Meer at:

+49 (0) 89 – 588 04 36 – 0

In recent years lexoro has specialised in Computer Vision topics as an IT solution provider and is therefore the ideal partner to implement your projects with you.

State institutions, pharmaceutical companies, retailers or companies in the manufacturing industry: use cases and fields of application of artificial intelligence are almost everywhere.

If you believe that your company is ready to head into the next few years fit for the future, please contact us using the form below.

More information about SCAN-X from our previous campaigns:

With SCAN-X we provide solutions for our customers so that they can easily scan objects and digitally process and continue to use them automatically.

The lexoro Group is a full-service provider in the markets of the future: Machine / Deep Learning, Computer Vision, NLP, NLG, RPA, DevOps and IoT.

Our range of services is comprehensive and our teams work with a range of companies on a great number of innovative and exciting IT projects. At the same time we help those companies develop their own products and also support start-ups in future markets. We are always at the centre of people and their connection with excellent technological competence.

With precision, stability and versatile usability in digital worlds.

In recent years, lexoro has implemented a variety of computer vision applications with practical relevance to the scanning process.

Here we have developed high-precision image solutions (ToF, 3D camera systems) as well as end-user solutions in the iOS and Android segment (dual camera).

Always in the centre of it: the easy execution of the scans with smartphones.

Our development teams have now combined this knowledge and building blocks in one platform: our SCAN-X construction kit.

For the first time, we offer companies in all industries the opportunity to make state-of-the-art computer vision technology available for their own application ideas.

  • For automating specific business processes (design, production), maximizing customer satisfaction (virtual-try-on).
  • And the development of completely new business opportunities in augmented reality or 3D printing.

The SCAN-X solutions open up a whole new space for the use of digital power and innovation.

The application possibilities are unlimited. The following is an impulse overview to stimulate your own thinking in appropriate application contexts.