Loading...

Follow Cortexica on Feedspot

Continue with Google
Continue with Facebook
or

Valid

With all the doom and gloom around in the retail world at the moment, it can be easy to lose sight of the bright spots.

Evolution of retail and the physical space it occupies is happening at hyper speed because of the impact of technology.

Within this maelstrom there are hundreds of people and companies changing the way physical retail works, doing innovative, creative and (importantly) profitable work. From developers and owners to retailers to tech companies, they are shaping the retail experience of the future…

Cortexica

Cortexica Another use of AI will be in the world of visual analytics, such as facial recognition technology that allows retailers to personalise offers to individual customers, although the reality is that this will be done via the equally intrusive but slightly less creepy method of recognising your smartphone. In the visual world, Cortexica created an app for UK shopping centre REIT Hammerson which allows consumers to point their smartphone at any object, be it an item of clothing or an everyday object in a colour they like, and find a store that sells that item or something similar.

Read full article here

The post 34 Companies Reshaping Retail Real Estate appeared first on Computer Vision, Machine Learning for Image Recognition and Video Analytics.

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

The world’s first edge compute solution for Personal Protective Equipment ingress and real-time PPW monitoring

Personal Protective Equipment (PPE) is the equipment that protects the user against health or safety risks. Not wearing PPE dramatically increases the chances of injuries and in many situations also of financial losses due to fines for injuries and death of workers as well as contamination caused by not wearing gloves, hairnets, shoe covers, etc. This post describes the world’s first edge compute solution for PPE ingress and real-time PPE compliance monitoring.

AAEON partnering with Cortexica allows us to identify numerous real-world problems that can be addressed by applying machine vision methods. has been at the forefront of the machine vision revolution. We have reverse-engineered parts of the human visual cortex, which allowed us to develop a powerful image search engine nowadays widely used to solve many real-world problems. Some of the solutions to these problems are often very specific while others are widely applicable and have the potential to save lives. Probably the best example of this is our family of solutions designed for PPE ingress and real-time PPE compliance monitoring.

check out the full blog post here

The post Real-time PPE Monitoring on the Edge saves cost and protect lives appeared first on Computer Vision, Machine Learning for Image Recognition and Video Analytics.

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

I am happy to announce that we have successfully developed the World’s first AI edge camera powered by the very latest Google Tensor Processing Unit (TPU) accelerator. This is the fourth time in the last two months that we have pushed the edge computing to its very limits. In December last year, we have created two types of AI edge cameras that are powered by multiple Intel Myriad-X Vision Processing Units (VPUs). The first version was based on Raspberry Pi and the second one on UP Squared board. The video below shows a quick overview.

More recently, we upgraded the first version of the camera so it could run off a small solar panel. This basically solves all the problems related to infrastructure needs because this camera is completely plug-and-play, connected via 4g and running off solar energy or a battery pack that can be recharged. The video below shows a very quick demo from the initial tests.

The first Raspberry Pi version of this camera, whether solar or not, was based on two Intel Myriad-X VPUs each running a different convolutional neural network based on mobilenet-ssd-v2 in parallel at around 10FPS. Same models run on UP Squared with Myriad-X on a mini PCIe (AI Core X from UP Shop) much faster at around 40FPS but to be fair I have to say that Intel clearly says that the OpenVINO support for Raspberry Pi is currently experimental. Therefore, I expect the performance to improve in the future. We are probably not going to see anything near 40FPS but I hope to see at least 15FPS with future optimisations.

Google developed its own Tensor Processing Unit (TPU) ASIC designed from the ground up for machine learning. These TPUs were until now used mostly to accelerate cloud-based machine learning. Recently, Coral announced the Edge TPUand provided two different means of allowing developers to test it out. The first is a full dev board while the second one is a USB accelerator similar to Intel Neural Compute Stick with Myriad VPU. We ordered the USB accelerator and developed the World’s first AI Edge camera that is using the TPU. At this point the tools provided by Google are quite limited so converting the model was a bit of a challenge but other than that I am very impressed by the performance of this tiny, low-power device. To quickly test it out, we converted our model for the detection of personal protective equipment (PPE)to Tensorflow Lite format and then we needed to upload this to an online service that compiles the Lite model for TPU. The resulting 8-bit quantised model has only around 5MB in size and runs on Raspberry Pi at impressive 25-30FPS!

Google TPU Accelerator

Here are few pictures that show the Raspberry Pi 3B+, camera module, and the Google TPU. This camera can be powered either via USB cable or by PoE. We use the USB power for solar version and PoE for places with existing infrastructure.

CortexiCAM TPU Test - YouTube

If you have any questions, let me know and I will do my best to answer as soon as I can.

The post The World’s first AI Edge camera powered by Google TPU appeared first on Computer Vision, Machine Learning for Image Recognition and Video Analytics.

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 
The BSiF (British Safety Industry Federation) has announced Cortexica as one of the finalists for its Product Innovation Award with AI-PPE Compliant.

This award further highlights the importance of innovation for safety and health professionals across multiple industries.

About AI-PPE Compliant

The use of personal protective equipment (PPE) is growing as a result of expanding health and safety regulations and shifts in work place safety expectations.

It is required that PPE be regularly inspected for safety and compliance purposes. And when the amount of required PPE increases, so does the strain on administrative resources in terms of reporting compliance. Combine this with the growing need for companies to prove their health and safety compliance in the event of an incident and you can see why Cortexica has developed ‘AI-PPE Compliant’, an action recognition solution that identifies what people are wearing on-site, making sure the correct PPE is being worn for the environment they are working within.

About the 2019 BSiF Safety Awards

The BSIF Product Innovation, Service, & Safety Solution Awards promote the importance of innovation and underline the highest standards of excellence within occupational Safety and Health.

The awards

Product Innovation:
An award for products that are new, innovative and will contribute to improvements in occupational safety & health.

Service Awards:
An award for companies in the safety field offering exceptional and innovative service solutions. Voted for by the entrants’ own customers.

Safety Solution:
Based upon case studies submitted by members, these awards recognise major improvements in occupational safety created in 2017/2018.

To find out more about the BSiF awards visit: bsif.co.uk/bsif-awards

The post Cortexica: Finalist in the BSiF Product Innovation Award appeared first on Computer Vision, Machine Learning for Image Recognition and Video Analytics.

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 
Cortexica’s visual commerce app is now live – download for iOS and Android!

Cortexica is extremely happy to announce that we have launched our very own app allowing users to demo our AI-powered visual search and image recognition technology.

How does it work?

Cortexica Visual Commerce app allows users to interact with a range of our retail solutions to get an idea of how visual search could work for them before having to invest in the technology. This is not to have our app confused with an e-commerce or transactional app (we leave that to the experts), what we are offering instead is access to our vision AI technology to understand how it enhances your existing retail offering.

So, how does it work? Well first of all download the app in either iOS or Android. Then simply login to the app (we have a login system to avoid disappointment for anyone trying to use the app for a shopping spree). Once you’re in you can start your visually empowered shopping journey…

You start by simply taking a photo on your phone or uploading a picture from your library, you can also picture from the in-app gallery of images, and then use the following products:

FindSimilar

Crop pictures to find similar products across multiple categories

Shop the Look

Automatically identify multiple items within one image and search for similar items

Colour Analyser

Extract the dominant colours from any image, be it fashion, home, or purely for inspiration

FindSimilar

Take a picture, or select a gallery image – crop to highlight the product you would like to search, then select the category you’d like to search.

 

Do you like the pattern of the dress but perhaps you want to see if you can find something similar in a skirt? Well now you can, once you chose the item you’d like to search, you can select any category to find similar items to the original image.

Our system automatically bases the search on 50% colour and 50% pattern to find you the most accurate results. To further personalise your search you are also able to refine the results, now if you feel like you want to see more of the particular pattern or colour, you can!

You can then either add to favourites to keep track of all your favourite items, or you can search ‘More like this’ to test more of what the app has to offer (we’ll explain more about this in a second). On your own app of course, this is where your customers would take their items to checkout.

Shop the Look

Shop the look demonstrates a different type of search functionality – instead of having to manually crop the item you want to search –  our systems do it for you.

Upload or select a gallery image and we will automatically locate all of the clothing items in that one image. Select if you want to search Men or Womens clothing, then the app will locate all the fashion items within that image.

Our visual search technology works out what items are dresses, shirts, skirts, trousers etc.. to save you having to crop and further filter the image yourself. Instead, all you had to do is click the clothing item and the app will find you similar items in that same category.

Many of our customers deploy this tool in their online offering to enhance the user journey and remove the number of steps the user has to take to see their desired items.

You may have noticed the ‘more like this’ button on all search results and selected items…

We have incorporated this button into our offering to showcase the inspirational user journey that it creates. If you see an item you like more than your initial search item you can select that image to search for similar items instead.

When incorporated into your website or app this tool not only creates a more intuitive user journey but also overcomes barriers such ‘out of stock’, ‘unavailable sizing’ and ‘out of budget’,  as the system can automatically suggest visually similar items and removes these obstacles in the path to purchase.

Colour Analyser

Colour analyser shows you the dominant colours in an image and their RGB codes, from the top 3-10.

Why do I want to see this you may ask? Well, first of all, your customers can use colour analyser to decipher the dominant colours in the fashion or home inventory images they like to learn what they want for their own purchases.

Colour analyser also has a wealth of potential applications in back of house retail operations. We can now extract data about the most popular colours chosen by your users. This enables brands to understand their users’ preferences and creates insights into visual analytics that explain what your customers are buying.

We’ve found that this ever-increasing image first world, combined with the advancements of vision AI, has created brand new insights into customer preferences and unlocked analytics from visuals that were previously inaccessible to businesses… Watch this space to hear more about our retail operational products made possible by visual search technology.

Seeing is believing right?

Sometimes no matter how many times you explain something, it doesn’t quite compare to seeing it in action. So rather than continue to tell you all about putting your customers’ needs at the forefront of the user journey and taking them from search to sale in a few clicks, we’re letting you try it for yourself – Download in the App Store or on Google Play now:

    

We will be rolling out more features very soon, stay tuned!

The post Cortexica launches visual commerce app appeared first on Computer Vision, Machine Learning for Image Recognition and Video Analytics.

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

The current state of PPE Compliance

Since the 1992 Personal Protective Equipment and Work Regulations Act came into force, businesses working in high-risk environments and required to provide Personal Protective Equipment, have been exploring new ways to monitor for compliance in as an efficient a way as possible.

Manually monitoring for PPE compliance among the workforce is still one method employed by many businesses today. There are many issues with this, particularly that human eyes can make mistakes, lose concentration, or simply not be focusing on the right place at that crucial moment. The risks of failing to notice a missing or faulty piece of equipment are high. Not only this, but the work-intensive nature of manual monitoring takes valuable time away from the on-site managers who could be focusing on other important work for the company.

However, as technology has evolved, potential solutions have emerged to help cut the workload. Most businesses now use RFID (Radio Frequency Identification) tags to monitor employee PPE-compliance. RFID tags are a way of tracking protective items via smart barcodes and radio frequency technology. The tags transmit information via radio-waves to the computer, which in turn translates the data into digestible information for the on-sight manager. In short, the tags can track what and when PPE items are being used.

However, while RFID tags are a smart technology that can be utilised in any number of ways, they are limited because they can only tell half the PPE story. RFID may show what protective equipment is being used, but not if it is being used properly. This is essential. What’s more, RFIDs can get damaged or malfunction, and there is no way of knowing in real-time that there is an issue with the data they are transmitting. The next solution in the PPE-monitoring process has to not only monitor WHAT is being used, but HOW it is being used. This is where Vision AI comes to the fore.

How Vision AI can transform PPE compliance in 2019

At Cortexica, we know that AI and machine learning is the future in PPE-compliance. Our Vision AI solution, called AI-PPE Compliant, can be implemented into any CCTV camera and allows business to constantly monitor PPE-compliance within the working environment. Unlike RFID solutions which require the implementation of a whole new system, AI-PPE compliant enhances existing systems, taking existing CCTV cameras and making them ‘smart’, with very little investment and integration. It flags in real-time not only if equipment is missing, but also if it is being utilised incorrectly. This means that the on-site manager can address the problem straight away before it leads to injury. AI-PPE compliant is trained to recognise a multitude of protective equipment, which allows companies to keep their existing PPE and not have to purchase costly smart helmets or install sensors on equipment and employees. Instead it visually monitors and works in the background with no interruptions to workforce and processes.

The benefit of this is two-fold. First of all, the technology is far more accurate in identifying potential issues, and because it can flag them in real-time, it drastically reduces the risk of injury in the workplace. Secondly, it collects, analyses, and processes the data far quicker than if it was a manual process, allowing those responsible for the safety of their employees to focus on new strategic and innovative ways to improve workplace safety.

In fact, recent research we undertook to inform our white paper: “Overcoming the barriers of PPE using Artificial Intelligence” showed that 81% of health and safety managers believed that AI PPE solutions would increase the speed and accuracy of detecting potential dangers, while 64% are looking to invest in AI continuous monitoring solutions for their business in the near future.

What will be the challenges to implementing AI PPE solutions in 2019?

While AI and machine learning are somewhat of a buzzword among businesses and the media, it is important for businesses to realise that AI is not one all-encompassing and overarching solution. It is a constantly evolving discipline.

As such, the real challenge for businesses looking to incorporate AI solutions in 2019 will be collecting the right data to train the system. Cameras need to be in the right place and capture the right data to be effective. With PPE-requirements varied across the different working environments, there is never an off the shelf solution. For instance, the data required to train cameras to identify PPE non-compliance on a stormy oil-rig in the North Sea is very different to a perfectly controlled and well-lit environment inside a pharmaceutical lab.

However, in 2018 we made great progress in training the systems in a variety of different environments. As an ever-evolving discipline, AI by its nature is always learning new tricks. Through tackling different PPE-related briefs (whether they be in construction, mining or manufacturing), our machine learning technology was always exposed to new data – constantly evolving to become quicker and more intelligent than ever.

Due to our considerable work in the sector already, in 2019 AI-PPE Compliantwill be able to  learn and adapt to each unique new brief quicker than anything else on the market, whilst still returning the best results for businesses across many industries.

Related Links

The post The challenges and solutions for PPE-compliance in 2019 appeared first on Computer Vision, Machine Learning for Image Recognition and Video Analytics.

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 
Cortexica by Sacha Lowenthal - 3M ago

If your business operates in construction, mining, manufacturing, oil and gas or any environment dealing with industrial machinery and volatile conditions, your workers are putting themselves at risk to deliver for the business. And while health and safety professional are already aware of the responsibility to keep employees as safe as possible – accidents can still happen, and your employees are not the only ones at risk if they get injured. The implications for your business could be immense.

We know that accidents happen and that humans make mistakes. It is impossible to eliminate the risk of accidents entirely. In fact, according to our data, the average business operating in high-risk environments had to deal with 27 non-fatal injuries that led to days off work in the last year. And nearly half (47%) of these were a result of human error.

While accidents are unavoidable, you can take steps to drastically reduce the risk of resulting injuries by ensuring that all workers wear the correct Personal Protective Equipment (PPE). Our own figures suggest that almost 3 in 10 workplace injuries are as a direct result of PPE failure or non-compliance, and these could be eliminated through proper monitoring and alerts. And if you look at the industry as a whole, hse.gov’s latest data shows that 16,746 accidents in the last year could have been prevented through the correct implementation of PPE in the workplace.

Not only could these casualties have been avoided, but so could the extremely large costs to businesses that follow them. The cost of PPE related accidents in Construction alone was £122million in 2017/18, and that’s just in the UK.

These statistics show that businesses cannot drop the ball on PPE – and it is vital that they maintain as high compliance levels as they possibly can, as often as they can. It is therefore extremely surprising that most businesses (84%) still manually monitor employees’ PPE compliance.

So, how can businesses reduce the risk of PPE failure to protect their employees from harm and themselves from losses? Our talented team at Cortexica have developed an automated solution using Vision AI and machine learning to monitor PPE compliance. It is called AI-PPE Compliant.

What is AI-PPE Compliant? Product 1: Access Point

AI-PPE Compliant comprises of two products and can be integrated into any camera. The first of these products is called ‘Access Point’. This analyses each and every worker outside the entrance of the worksite, and only allows them through the entrance if they meet all the PPE requirements. Should they be missing a piece of equipment, they will be prevented from entering the site.

Product 2: Continuous monitoring

Nevertheless, ensuring employees enter the workplace sufficiently protected is only half the battle. Workers can take the equipment off once on site for any number of reasons. Areas can get hot, equipment can become heavy, it can be momentarily removed or forgotten to put back on.

Therefore, the second solution to AI-PPE Compliant is called ‘Continuous Monitoring’ – and it does exactly that. This product constantly monitors everyone within the site 24 hours a day for PPE compliance – without relying on manual checks that are prone to human error and external variables.

Should any ‘non-compliance’ be detected, the system will alert the on-site manager instantly through any form of communication. This way, non-compliance can be dealt with in real time and injuries can be prevented, saving companies in the UK hundreds of thousands, if not millions, of pounds.

 Not only this, butAI automation can save valuable time collecting and analysing data for those responsible for the safety of their employees. This allows them to focus on new strategic and innovative ways to improve workplace safety, instead of their time being taken up manually monitoring worksites.

If you’d like to find out more about overcoming the barriers of PPE compliance with artificial intelligence you can download the whitepaper, or view our AI-PPE Compliant solutions. We will also be taking a closer look at the challenges and solutions for PPE-compliance in 2019 in our next blog post.

The post AI on-site: keeping workers safe appeared first on Computer Vision, Machine Learning for Image Recognition and Video Analytics.

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

This content is password protected. To view it please enter your password below:

Password:

The post Protected: draft blog for martin – for Stuart’s approval appeared first on Computer Vision, Machine Learning for Image Recognition and Video Analytics.

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

I am very excited to say that Cortexica Vision Systems have managed to develop a super fast AI Edge Camera that is capable of running multiple deep learning models in parallel and thus enable novel applications that would previously be out of reach.

In a short period of time Cortexica has become the first company to demonstrate the power of the very latest Intel Myriad-X VPU (Vision Processing Unit) at the recent IOTSWC conference that was held in Barcelona. We have presented our AI Safety solution running on our new product developed in partnership with UP and AAEON. Check out this interview and this webinar for more details.

Couple of months has passed since the IOTSWC and I am super excited to share with you yet another innovation , which I believe is the first of its kind in the world. Cortexica is very much hardware and software agnostic, which allows us to spot the emerging opportunities and act on them fast. Only couple of days ago Intel released OpenVINO R5 enabling

Raspberry Pi boards to be used as a host for Intel Neural Compute Sticks (NCS) that is leveraging the power of the Myriad-X VPU. Only couple of days since this release and Cortexica has managed to create the world’s first AI Edge Camera powered by two of these super low-power high-performance Myriad-X processors!

Read the full article and find out everything you need to know to replicate and create other exciting applications here

Introducing Myriad X: Unleashing AI at the Edge - YouTube

The post The World’s first AI Edge Camera powered by Raspberry Pi and two Intel Myriad-X VPUs appeared first on Computer Vision, Machine Learning for Image Recognition and Video Analytics.

Read Full Article
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Separate tags by commas
To access this feature, please upgrade your account.
Start your free month
Free Preview