HANGAR 51
Speed reading
Felicity Stredder spoke to AllRead Machine Learning Technologies,
one of 13 winners in the first stage of IAG’s Hangar 51 global
innovation programme.
This year, IAG launched
its fourth Hangar 51
global innovation
programme, which offers
a select few innovative
companies ten weeks of
personalised mentoring, the
opportunity to pilot their
disruptive technologies on
a large scale and the chance
to receive investment from
the IAG venture fund,
as well as fast-tracked
commercial opportunities
within the group.
The project is growing
in popularity with every
year. “It’s harder to get into
Hangar 51 than into Harvard,”
asserts IAG’s Head of Global
Innovation, Dupsy Abiola.
Indeed, more than 450
applications were received
from over 50 countries this
year, whittled down to just
36 applicants across seven
categories for the initial
pitch day, held in Madrid
this September. From there, a
total of 13 entrants were put
through to the accelerator
programme, which began on
September 30 and ended on
December 6, 2019.
The cargo category
Winning in the Future
Cargo Logistics category was
AllRead Machine Learning
Technologies, which was
paired with IAG Cargo as its
mentor for the accelerator.
COO Adriaan Landman and
CEO Miguel Silva-Constenla
presented the company’s
concept at the pitch day:
computer vision software that
identifies and reads any kind
of text or symbols found in
industrial contexts, be they
serial numbers, barcodes or
expiry dates, and converts this
into big data.
Says Silva-Constenla: “Data
team has been working together within the venture builder The
Collider, a technology transfer focused programme by Mobile
World Capital, for six months, leading to the founding of AllRead
MLT in March this year. The three engineers are the inventors of
the technology, which is the result of five years of investigation
inside the Computer Vision Center of Catalonia (CVC). The
technology and its patent have been transferred with exclusive
rights to AllRead MLT.”
Developing the product
Landman relates how the computer vision software came to
be developed. “The inventors of the technology are scientific
investigators. They worked in a group that specialised in ‘robust
reading’, which refers to the research area dealing with the
interpretation of written communication in unconstrained
settings, and had been mandated by a Spanish gas company to
help them develop a solution to ‘read’ analogue gas meters. Once
they reached 98% reading accuracy, the team worked to apply the
technology in new contexts,” he explains.
“Logistics is significantly affected by the costs of tracking and
digitising text on products, vehicles or load devices, and AllRead
MLT saw a potential application for their computer vision
software technology.”
Since then, the product has gone from potential to reality.
“Today, AllRead MLT offers an (online to offline) machine
learning based software to process images (photos and videos)
taken by mobile or fixed cameras, and to spot and digitise
any type of alphanumeric text in operational environments,”
Landman relates. “Through its agile and easy implementation,
it streamlines and simplifies data extraction processes, reducing
repetitive manual tasks and allowing immediate processing of
is the new oil. We want to give
access in a frictionless way to
the valuable data appearing as
text in the supply chains of our
clients, transforming the raw
data into crystal clear data.”
The technology was
championed as a way to
reduce monotonous tasks,
saving time and money in
labour, while also enhancing
accuracy, even in challenging
operational environments,
such as when target text is
moving or obscured.
The team at AllRead
is a combination of two
experienced entrepreneurs,
Landman and Silva-Constenla,
with international track
records in start-ups and
corporate companies, and
three PhD-qualified senior
computer vision and machine
learning engineers. With one
Greek, a Dutchman and three
Spaniards, the team is very
international, speaking ten
languages between them.
Landman explains how the
company was founded. “The
12 December 2019 www.airlogisticsinternational.com
/www.airlogisticsinternational.com