ev.lib

EV-lib

EV Lib is a complete embedded vision software library based on deep learning and AI with functionalities for people, vehicle and object detection, identification tracking, and 3D pose estimation.

Keywords:
Embedded Vision Library, Deep Learning, Machine vision, Commercial Exploitation, Business Plan

Vision

At Irida Labs, we are developing a machine vision library for embedded devices called EV-Lib – Embedded Vision Library. This is a globally competitive technology and product, as can be seen from the description of the company’s activities and export activity to date.

The library is based on machine and deep learning technologies where significant research and development has been, and continues to be, carried out by Irida in the past years in the framework of European H2020 research projects. Additionally, Irida Labs is also investing further private resources to develop the EV-Lib technology and product by deepening the fields of machine learning.

The results of this investment and development to date are very encouraging both on a research and commercial level. At the research level, in addition to the patents filed, highly competitive results have been announced. In the context of the EV-Lib project, the company will expand this product both technologically, in order to remain competitive at a global level, and commercially by targeting new markets.

Functionalities

EVLib offers wide flexibility for specific object detection like food recognition, product catalogues recognition and numerous other applications. Pre-trained models can be easily extended for custom object detection and tracking and can be customized to any computer vision application using scenario-specific data.

Product Recognition & Analytics
Food Recognition
Shelf Monitoring
Vehicle Classification & Tracking
People Counting
Object Pose Estimation

Goals and Objectives

1.

The continuous technological evolution of the EV-Lib library (by integrating new functionalities such as car detection and classification, detection and classification of different – custom – objects) to remain at the cutting edge of technology while remaining competitive on a global level.

2.

The preparation of a feasibility study and market analysis for the EV-Lib product and the exploration of a business model in relation to its further commercial exploitation. This development concerns the study of markets and scenarios where this technology can be used.

3.

The development of a feasibility and market analysis study for the EV-Lib product in relation to its further commercial exploitation in different geographical markets (e.g. America, China, Korea, Europe, India).

4.

The strengthening of intellectual property with patents at the USPTO, EPO and OBI.
Participating in international exhibitions and holding individual B2B meetings to commercialize EV-Lib.

Methodology - Project Phases

The EV-Lib project is a continuation of the IRIDA business plan and will build on existing research results. The implementation methodology aims at the continuous technological development of the EV-Lib embedded vision library in order to keep it on the cutting edge of the global market, its commercial exploitation through IP rights and opening up to new markets through feasibility studies, as well as participation in international exhibitions as exhibitors and visitors in order to demonstrate the technology and search for potential partners and customers.

Two distinct and interlinked phases are included:

Phase 1: Continuous Technological Development

Intellectual Property Registration

Technology Innovation Services

Phase 2: Feasibility Study & Commercialisation

International Exhibitions

Commercial Exploitation of the Systems

Work Packages

Gantt EVLib

Results

Smart Retail - Shelf interaction demo deployment

This system provides alarms for each product pickup (each time a product is picked up from the shelf). To help the user evaluate the performance, the number of product receipts is also displayed in the plugin’s web interface. So in case a person picks up ten (10) products, the app will trigger a ‘Product Interactions’ alarm and the counter will be equal to 10. So when staff are checking price tags, placing products on shelves or customers are touching/reaching products there is no alarm.

shelf interaction EVlib
Left : Store staff checking prices / stickers and moving a product from one shelf to another. This fact does not raise an alarm. Right: a customer takes a product from the shelf and an alarm is generated

Parcel identification demo deployment

User can track the counters values directly from the main plugin web interface. The metadata provides information about line crossings (incoming and outgoing parcels) as well as the presence of frames as shown in the figure below:

Percv.ai Parcel Analytics
Parcel Metadata Visualisation

Dissemination

Irida Labs, to further commercialize the EV Lib technology, has chosen to participate in leading international exhibitions related to embedded vision, machine vision, embedded electronics, and the Internet of Things. This participation can be done in two different ways. Firstly, in the form of a direct exhibitor (direct exhibitor), i.e., with a separate stand and presence, or in collaboration with internationally recognized companies as “guests” in their own space presenting joint products and demos. The following paragraphs present examples of participation in both ways in international exhibitions in Europe.

Participation in the VISION 2021 exhibition

Irida Labs in October 2021 (October 4 to 8, 2021) participated as a direct exhibitor in the largest machine vision exhibition in Europe, VISION 2021, which takes place in Stuttgart.

Irida Labs participated with its booth measuring 3×6 square meters, open on two sides. The company’s team consisted of Mr. Anastasiou, Papakonstantinopoulos, and Tsagaris.

In the context of this exhibition, the Warehouse Management solution based on the EV Lib bookcase was presented.

Participation in the Embedded World 2022 exhibition

Visit us at Embedded World 2022

Irida Labs has participated several times in the Embedded World exhibition in Nuremberg. The show was initially scheduled for the beginning of March 2022, but due to the pandemic, it was decided to move it to June 2022 (June 21 to 23, 2022). Irida Labs, having already planned its participation, has paid the cost of renting the space for the stand where the EV Lib results will be presented.
The company’s participation has already been announced on the website.

Participation in the digital exhibitions

During the pandemic and to continuously promote the EV Lib library and the project results, Irida Labs also tried alternative forms of promotion, such as participation in digital exhibitions such as the Embedded Vision Summit 2020 (15 to 25 September 2020). Irida Labs has previously participated in the physical equivalent show.
In the same context, Irida Labs participated remotely in Edge AI Summit 2020 (November 18-19, 2020) and Space4 (June 22-23, 2021). In general, the experience of purely digital exhibitions could have been more positive, and for this reason, future planning focuses on physical presentations and meetings with customers when possible.

Marketing

As part of the project implemented by Irida Labs, various publicity and digital marketing actions have been carried out. The publicity actions included creating promotional material such as t-shirts and business cards, mainly during the participation of company members in trade fairs as described in detail in the relevant deliverable. In particular, the t-shirts with the Irida Labs logo were used at all events and exhibitions, such as the VISION 2021 Renesas FAE meeting, as seen in the relevant photos of deliverable P14.In the same context, Irida Labs participated remotely in Edge AI Summit 2020 (November 18-19, 2020) and Space4 (June 22-23, 2021). In general, the experience of purely digital exhibitions could have been more positive, and for this reason, future planning focuses on physical presentations and meetings with customers when possible.

Webinars

In promoting the EV Lib library, Irida Labs co-sponsored the webinar organized on March 24, 2021, by Vision System Design magazine entitled “Deep Learning for Embedded Computer Vision: An Introduction,” presented by Jeff Bier. Three hundred four participants registered and attended the webinar, and as a result of the sponsorship, Irida Labs had access to their emails registered in our internal database.
In addition, Irida Labs created its content as a webinar.

Other publicity actions

Sponsored Webinar AI in Machine Vision
Sponsored Webinar AI in Machine Vision

A tool used to promote the project’s results was advertising through the LinkedIn platform. In particular, two campaigns were carried out for people counting and palletizing system applications for Industry 4.0. In addition, Google Ads services were used for a short time.

In the context of promoting the EV Lib library, Irida Labs co-sponsored the webinar organized on March 24, 2021, by Vision System Design magazine entitled “Deep Learning for Embedded Computer Vision: An Introduction.” A second sponsorship of the Imaging and Vision Europe webcast took place in December 2020

An alternative form of digital marketing and promotion of EV Lib was through mass emailing to Site Media’s Engineerlive.com magazine database/readers.
To support these publicity and digital marketing actions, images or videos from platforms such as Shutterstock, Istock, and Fiver were also used to create appropriate promotional material.

Contact





    Funding Details

    This project has been co‐financed by the European Regional Development Fund of the European Union and Greek national funds through the Operational Program Competitiveness, Entrepreneurship and Innovation, under the call RESEARCH – CREATE – INNOVATE (project code:Τ2EDK-03212)

    Skip to content