The machines have eyes
Machine Eye is giving the power of sight to industrial equipment
Welcome to another PreSeed Now. Today’s startup is bringing computer vision to industrial machines, showing that there’s far more to the future of autonomous vehicles than cars, trucks, and those funny little minibuses you see driving around university campuses. Scroll down to read all about Machine Eye, along with details of two startups currently raising rounds.
A special hello if you’re a new subscriber since last Thursday’s edition. Don’t miss out on last week’s profiled startups: NeuroVirt and Enzai. The PreSeed Now homepage is already starting to look like I envisioned it when I started this newsletter - a growing collection of startups worth exploring if you want to understand the breadth of pre-seed activity in the UK. There’s plenty more to come!
Speaking of which, I’m keeping the newsletter fully free to read for the rest of this week and then starting the transition to making new articles restricted to paying subscribers from next week. Free subscribers will still get a monthly roundup email. Want to upgrade your free subscription? Hit ‘edit’ next to PreSeed Now in your Substack settings. Thanks to everyone who has already done this!
As ever, if you have any comments, questions, suggestions, or you know a great early-stage startup I should be covering, please get in touch. just email firstname.lastname@example.org
💷 Now raising
Early-stage startups raising money right now…
My2be counts Box and Nextdoor among the customers for its software designed to easily start, run, and scale company-wide mentoring programmes. The startup is built on the premise that these programmes can be beneficial to businesses and employees alike, but they can be difficult to run at scale in large organisations without dedicated tools.
Bootstrapped to date by Manchester-based co-founders Adam Mitcheson and Damien Shiells, the startup is raising its first equity round to accelerate its growth through selling into employers looking to boost employee diversity, talent development, and retention.
Audemic is targeting academia and business with an easier and more efficient way to consume academic papers: via audio. The Durham- and Madrid-based team has developed an MVP app with features designed to help users quickly consume the parts of a paper they need to understand, or listen to it all. This is particularly useful if your research involves hundreds of papers.
CEO Joshua Mitcham says Audemic has attracted interest from academia and the pharmaceutical sector. Mitcham says that after closing a £90,000 pre-seed round in late April, the startup is currently raising a bridge round of £150,000 to double down on product development and sales ahead of a planned seed round.
Machine Eye is giving the power of sight to industrial machines
From quarries to farms, autonomous vehicles have potential far beyond the highway
The hype around self-driving cars has subsided since the days when seemingly every manufacturer boldly predicted they would have autonomous vehicles on the road by the early 2020s. But still, most of the attention the field receives focuses on cars and trucks. But what if there are greater near-term gains to be had from focusing on bringing autonomy to other kinds of vehicles?
Machine Eye is a Belfast-based startup developing what it describes as “a vision based operating system for the control of off-highway machines”. In practice, this means helping the likes of forklift trucks and loading shovels to operate more safely and efficiently in spaces like warehouses or construction sites with the aid of computer vision.
Founder Brendan Digney explains: “A self-driving car going down the road has a lot of things to consider. It has to consider other traffic, it has to consider navigation, it has to consider people, it has to consider the rules of the road. And there's a lot of really good work being done there. However, as a society, how close are we to accepting that [a car] doesn't have a driver in it at 70 miles an hour?
Industrial fields like quarrying or agriculture have a lot fewer variables to worry about, not to mention a strong commercial incentive to be as automated and efficient as possible.
“These machines are working often in a large quarry where there's no members of the public. They're working in a field where it's a very regimented environment. It's within the boundary of four edges or four fences, moving back and forth and progressing at a known pace,” says Digney. “So in terms of readiness for autonomy, the off-highway sector is a lot closer to that reality because you don't have many of the same risks associated with an autonomous machine as you do on the public highway.”
And in some cases, Digney says, the off-highway machines we see today already have some level of autonomy.
“If you travel down a motorway today and you see a tractor working in a large field… that tractor is to a certain extent semi-autonomous or not directly under the operator's control. It will have a GPS receiver or an RTK receiver on the front, it'll be guided by an external source. It'll be running a program which decides its guidance.
“There will be an operator on the machine, but… they’re effectively just being an observer to ensure there's no risk, there's no people, no objects in front of the machine and maybe turn it at the headland and make sure everything's running as it should.
“So if you take that scenario, there's a machine that's running quite a high level… ‘autonomy’ is not the word, but a high level of known human control… in a contained area. Really, what's left to add? That is the ability to sense what's going on around it”.
And that’s where Machine Eye comes in, using computer vision to provide ‘sight’ for machines. While autonomous operation is on the roadmap, at present the startup is focused on driver-assist features that help workplaces remain safe and legally compliant.
“We're trying to give drivers that extra support ,trying to… augment their own situational awareness and their own capabilities. And to give business operators a technology that can start to support them, start to assist them in better managing and better operating the businesses” says Digney.
The eyes of the machine
If computer vision has passed you by, it’s a field of A.I. and machine learning that allows computers to ‘see’ almost like humans do.
“As I look around a room here now, I'm instantly saying ‘that's a Coke tin, that's a coffee cup… that's a computer, that's a shelf’. We don't think about that, we do that identification of an object and classification of an object [in our brains] instantly. We can without thinking establish whether a person is moving towards us away from us, left to right, based on the change of perspective and the scale of the image.
“That's something that we [as humans] don't think about. What computer vision is, is digitally recreating that ability... So what we're effectively doing with Machine Eye is trying to give those machines that same capability that we have through sight to identify objects, to understand what the objects are and to understand how the objects interact with machines”.
Computer-vision based autonomous vehicle systems have something of a controversial reputation. While Elon Musk has gone all-in on using cameras to provide autonomous features in Tesla cars, many working in the field prefer the safety of using multiple systems to monitor what’s happening around a vehicle, such as Lidar to accurately map the surroundings in real-time 3D.
But Digney says computer vision is a better solution for the environments Machine Eye is targeting, because the conditions can change so quickly.
“A dry day can suddenly become overcast, it can suddenly become wet, so you have moisture in there - a lot of water in the air. You can have dust and particulate, so suddenly you have solid matter suspended in the air… Some of those conditions can cause other technologies to struggle, where generally with vision, if we as humans can see it, a vision-based system can see. So you have a broader operating window where environmental conditions, maybe aren't pristine or optimal”.
From R&D to delivery
Digney founded Machine Eye following his graduation from a Masters degree in electrical engineering at Queen’s University Belfast. Having his eyes opened to the possibilities of computer vision during the course, he tapped into support from the university and the entrepreneurial community in the city to get his company off the ground. This included taking part in Ignite NI’s Propel pre-accelerator programme in 2020.
Today, the seven-person Machine Eye team is split between Belfast and Southampton in the UK, and Galway in the Republic of Ireland. “We've a really specialist team of really great people,” says Digney. “Some have joined us because they worked with us on grant-funded projects or research projects… others, it's been personal introductions. We have a really close knit team.. people who are really sure of the vision that we have - no pun intended.”
The company is currently making a gradual transition “from an R&D mentality to a delivery mentality”, as Digney puts it, although he says this will take time. When I was arranging a time to talk to him, we had to delay by a few weeks as he was out on multiple site visits to locations where businesses are already using Machine Eye technology.
“We’re lightly touching the market… We're on site, we're out with these machines, we're with people who use them or back [in the office] doing algorithm development, annotation, refinement… It's a really a broad church, what we do at the moment.”
Digney stresses that Machine Eye is very much focused on the situational awareness element of future autonomous machines, and other companies will provide other parts of the array of technology required. He explains that everything from motor design, to sensor placement, to planning the environment in which the machines operate will all play a part in the transition to autonomy.
“What you're going to see is a consortium of businesses coming together…. we would hope that not so many years as this becomes more prevalent that we all have an important role to play in that.. and in that overall package that results in autonomy off-highway.”
Digney says Machine Eye has been largely funded to date by a mix of Invest Northern Ireland grants, Ignite NI’s equity-free funding, support from Queen’s University, and pitching competitions. It also raised what he describes as a “small” pre-seed round last year to grow its team, and is likely to raise another round soon.
That’s all for now
PreSeed Now will be back in your inbox on Thursday. Know a team building something great? Please put me in touch! Just drop me a line: email@example.com
Were you forwarded this newsletter, or maybe you found it on the web? Become a subscriber to get the newsletter in your inbox every Tuesday and Thursday. Just enter your email address below.