Computer Vision

Bringing intelligence to vision

AI automates your visual tasks.

Following the giant leaps made in recent years in Artificial Intelligence technology and its uses, Inetum has taken on the subject with its many applications, including vision.

Powered by mixed models combining the reliability of classical computing with the computational power of the latest models of artificial neural networks (Deep Learning), our Computer Vision offer aims to automate the various vision tasks conducted in your business lines.

Available both in the Cloud and on-premise, our bank of visual detection algorithms and its excellent modularity make it possible to find the most suitable technological response to a specific need for automation in a vast array of fields, ranging from industry to character recognition, including the Smart City.​

Optimized visual verification thanks to automation​

Support your operators with an intelligent vision system. Computer vision automates repetitive, lengthy and hazardous verification and inspection tasks with little added value. Quality control in an industrial environment, document remediation and identification in banking and insurance, and the production of traffic statistics for a road link are all examples of tasks where an operator can be supported by an intelligent vision system that can support them with their work and get more reliable results more quickly. This process automation saves a lot of time and resources over the medium to long term.​

 End-to-end AI expertise to meet your visual automation needs

Read more

Support in complex and cross-developmental situations (creating a database, recovering pre-existing video streams, installing recording equipment, etc.)​

Read more

Sound knowledge of AI design and operating environments

Read more

From idea generation to innovation​

We support you in your computer-vision project from start to finish. ​

​Inetum's Computer Vision offer supports you from idea generation to production and runs the vision solution once it has been identified and developed. Our support phases include:​

  • Workshop on idea generation to innovation including the detection of elements in your business units that can be automated; ​
  • Tests and feasibility studies to validate the elements identified beforehand; ​
  • Development of the solution; ​
  • Repetition with the customer teams if necessary; ​
  • Support from our industrial partners in producing the solution of additional devices (industrial conveyors, special IP-standard cameras, IS interfacing, etc.) if needed;​
  • Go-live in the Cloud or in a local environment, according to your needs;​
  • Agile project management tailored to your business

Offers dedicated to your vision

Our catalogue of solutions covers a wide range of technological needs.​ See our catalogue of computer vision offers to find the solution that best meets your needs:​

Visual inspection – Industry 4.0 

Visual inspection is a key step in the industrial production process. Whether it is a question of checking the compliance of a production campaign with an internal standard or an industry standard, the partial or final review of a product is still an essential part of a production process that meets the applicable quality requirements. The automation of this step, which already exists in many industrial sectors, is experiencing further expansion with the development of visual artificial-intelligence technologies. Combining the reliability of the machine with the expertise of the human eye, this is the challenge for deep learning by offering a whole new range of detection programs supported by the progress made in the sub-domain of artificial neural networks. Discover the expertise of our vision teams on this subject. ​

Smart City​

Giving cities the technological tools of tomorrow: it is with this in mind that Inetum is offering its vision expertise to local authorities. The Smart City faces many economic, social and environmental challenges, and must use the most effective technological means to achieve its objectives in terms of urban development, traffic regulation and energy saving. ​

Investing and innovating to save and manage better – this is the mindset behind the Vision Smart City offer. Examples of projects aimed at optimising urban-management projects conducted using Inetum vision tools are the ability to trigger and automate sequences of actions on the basis of detected events (traffic jam, incident on the track, etc.) or to create a reliable, anonymized and precise statistical database on the use of the various road and pedestrian links of a city based on a pre-existing video system.​

Document recognition

The objective of the Inetum Document Recognition offer is to use digital vision tools to automate the administrative tasks and workflows involved in the processing of the captured information. Identifying the stock or flow sequences processed manually, developing the relevant algorithmic bricks and creating a control interface for monitoring operations are the different stages of a document-processing automation project. OCR, LAD/RAD, custom AI models for complex documents – our range of products and services for document-related purposes covers a broad spectrum of practical scenarios in this area.​

Tailor-made vision project

Do you have a specific visual-detection project that does not match the practical scenarios listed in our offers? ​

Our teams have the skills and agility to build a tailor-made vision program, perfectly suited to your constraints and your field and/or software requirements. Contact us to make an appointment with our teams on this subject in order to quickly obtain an initial project qualification meeting.

Leading industrial partners for relevant solutions

A Computer Vision ecosystem at the service of your projects.​

For its Computer Vision offer, Inetum's innovation team has forged various partnerships and collaborates with key players in the design and rollout of vision systems. As a member of the Collectif Continuité Numérique (digital continuity collective), which groups together the manufacturers ATS, Festo, ifm electronic, Phoenix Contact, SICK, and the software developer SAP, Inetum is able to address and manage relevant partners with complementary resources to the different types of vision projects. We support you from start to finish in choosing the conveyors, insulation housings, lighting, cameras, pneumatic displacement systems, electronics, PLCs and software interfaces needed for your project.​

Our client references​

A subsidiary of the Bostik industrial glue production group wanted to automate the quality control of one of its glue-pack production lines at a pilot site in the Netherlands. Until recently, the packs were visually inspected one by one by an operator, who was also responsible for retrieving those with defects.​ ​

The solution developed consists of a computer vision system intended to inspect each of the glue packs by detecting any defects in under one second. ​ ​

The chosen technology uses deep-learning algorithms (Tensorflow API detection).

IT-CE manages the IT systems of BPCE (banking group). The group has to deal with the management of millions of unstructured data collected since its digitalization. The challenge was to deploy a bespoke computer vision algorithm package to read, interpret and classify the documents (mainly passports and identity cards) in order to sort them into categories and comply with the latest regulatory requirements of the banking sector.

The municipality of Roeslare (Belgium) wanted to deploy a system capable of detecting traffic jams on a main street in town. The objective of the system was to alert the cyclists ahead of traffic so that they could be given a suggested alternative route, using a display panel indicating the incident. ​ ​

The system was configured to send an alarm whenever a "traffic jam" situation is detected by the camera. The alarm was then linked to a display system broadcasting a visual warning message upstream of the traffic. Another system was deployed to count the bikes that followed the warning advice and actually changed route. They also wanted the system to detect the number of bikes that actually follow the recommended route displayed.