Showing posts with label computer vision. Show all posts
Showing posts with label computer vision. Show all posts

Friday, 21 August 2020

So I heard you can do Computer Vision at 30FPS; I can do 1000.

- Akash James



And there was a man, in a cave, held captive and hooked up to an electromagnet plunged deep in his chest. Hammering his way through, quite literally, Stark, built his initial Arc Reactor and Mark 1 Iron Man suit, using nothing but a bucket of scrap and modern, tactical, self-guiding, explosive payload-carrying arrows, ergo missiles. Over-did it, didn’t I? Mesmerizing to most, the primitive propulsion system for un-guided flight and rudimentary weapons were not striking to engineers like us.

Stark kept going on, adding new capabilities to his armour, reaching peak performance with the Model Prime and finally calling it a day with the Mark 85. (More like Captain Marvel blasted him in Civil War 2 or the Gauntlet irradiated him, based on the cinematic or comic universe you prefer).

Just like arguably the best science-fiction-based inventor, I never stop with my creations and continue over-hauling for higher performance, ’cause I know that there will always be a higher ascension level to reach.

Computer Vision is a field with rapid progress; new techniques and higher accuracy coming out from various developers across the planet. Machines now have human-like perception capabilities, thanks to Deep Learning; with the ability to not only understand and derive information from digital image media but also create images from scratch with nothing but 0’s and 1's.

How did it begin?

Time and again, the higher tech-deities bring me at a point in this space-time continuum where I am faced with a conundrum. My team and I, back in our final year of college, were building a smart wearable for people with impaired vision, an AI-enabled extension of sorts to help the user with recognizing objects, recognizing people, and performing Optical Character Recognition; we called it Oculus. In all honesty, we did not rip it off from Facebook’s, Oculus Rift VR Headset and it was purely coincidental. The AI Engine was comprised of a multitude of classifiers, object detectors and image captioning neural networks running with TensorFlow and Python. With my simpleton knowledge of writing optimized code, everything was stacked sequentially, not allowing us to derive results in real-time, which was an absolute necessity of our wearable. Merely by running the entire stack on the GPU and using concurrent processes, I was able to achieve 30fps and derive real-time results.

Thus, this began my journey of being fast — real fast.

Ratcheting my way through

Fast forward two years to the present, I currently work as an AI Architect at Integration Wizards. My work predominantly revolves around creating a digital manifestation of the architecture I come up with for our flagship product — IRIS

Wondering what exactly IRIS does? (being Deadpool and breaking the 4th wall) To give you a gist, IRIS is a Computer Vision platform which provides our customers with the ability to quickly deploy solutions that monitor and detect violations. People counting and tracking with demographics, adherence to safety gear usage, person utilization, detection of fire, automatic number plate recognition and document text extraction are some of the features that come out-of-the-box. 

Typically, IRIS plugs into existing CCTV networks, rendering previously non-smart recording networks into real-time analytical entities. IRIS uses Deep Learning for it’s AI Engine but the architecture of the pipeline and the neural networks has seen many changes. My first notable architecture involved web technologies, like Flask and Gunicorn, to create APIs, that my worker threads could utilize. This ensured that the GPU was utilized in a better manner. However, this turned out to be moot when a large number of streams were to be processed.

The two primary hindrances were the API based architecture being a bottleneck under higher loads and the Object detection neural networks being heavy. For this, I needed something better, a better queue and processing architecture along with faster neural nets. Googling and surfing Reddit for a couple of days, I came across Apache Kafka, a publisher-subscriber message queue that is used for high data traffic. We retro-fit the architecture to push several thousand images per second from the CCTVs to the neural networks to achieve our analytical information. We devised another object detection model that was anchor-less and ran faster while retaining performance. Of course, the benchmark was against the infamous COCO dataset.
This increased our processing capability close to 200 fps on a single GPU.

The Turning point

Yes, you guessed it, I didn’t stop there. I knew that there was much more fire-power I could get; accessible but hidden in the trenches of Tensor cores and C++ (such a spoiler). The deities were calling me and my urge to find something better kept me burning the midnight fuel. And then, the pandemic happened.


WHO declared COVID-19 a global emergency — it ravaged through multiple countries and fear was being pushed down people’s throats; most offices transitioned into an indefinite work-from-home status and India imposed the world’s largest lockdown. Wearing masks and Social distancing was the new norm and everybody feared another Spanish flu of the 1900s. 

As an organization, we work with AI to be an extension of man, helping the human race to be better. Usage of face masks and social distancing needed enforcement and what better way to do it than with AI? Our stars aligned, the goals matched and we knew what we needed to build. The solution had to be light-weight and fast enough to run on low-end hardware or run on large HPC machines to analyze hundreds of CCTV cameras at once. For this, we needed an efficient pipeline and highly optimized models.

Hitting 1000 with Mask Detection and Social Distancing Enforcement

By now, I had a few tricks up my sleeve. IRIS’ pipeline now harnesses elements of GStreamer, which is an open-source, highly optimized, image/video media processing tool. TensorRT is something we used to speed up our neural networks on NVIDIA’s GPUs to properly utilize every ounce of performance we could push out. The entire pipeline is written with C++ with CUDA enabled code to parallelize operations. Finally, light-weight models — the person detector uses a smaller ResNet-like backbone and our Face Detector is just 999 kilobytes in size with a 95% result on the WiderFace dataset. Our person detector and Face Detector are INT8 and FP16 quantized making them much faster. With quantization and entire processing pipeline running on the GPU, amalgamating these together, IRIS’ new and shiny COVID-19 Enforcer ran at 1000 fps at peak performance for Social Distancing and 800fps for both Social Distancing and Mask Detection.

This allows us to deploy IRIS on smaller embedded devices to provide a cost-effective solution for retail-chains and stand-alone stores while letting us utilize multi-GPU setups to run on warehouses, shopping malls and city-wide CCTV networks making it easier to comply with and deny the spread of infection.

So what’s next?

I am not done. Achieving one milestone allows me to mark a bigger and better goal. Artificial Intelligence is in its infancy and being at the forefront of making it commercially viable and available in all markets, especially India has been mine and my organization’s vision. The endgame is to have AI for all, where people, be it developers or business-owners, have the ability to quickly design and deploy their own pipelines. 

IRIS aims at being a platform to precisely empower individuals with that, with the intention to democratize Artificial Intelligence, making it not a luxury for the few, rather a commodity for all. 

Chiselling AI agents to be the best tool that man has ever known will be our goal, paving the future with a legion of Intelligent agents, not making the world cold, but making us a smarter race. Ain’t nobody creating Ultron!

Thursday, 2 July 2020

Enhancing Workplace Health & Safety Using Computer Vision

Subhash Sharma 


Although health and safety at workplaces have improved over the years, yet the UK continues to have a large number of workplace accidents. The number of accidents resulting in injury, or in some cases, even death is quite high. Many of these accidents can be avoided, and AI-based computer vision can play a significant role in cutting down these accidents. 
Health and Safety Statistics. Key figures for Great Britain (2018/19)
  • 1.4 million working people suffering from a work-related illness
  • 2,526 mesothelioma deaths due to past asbestos exposures (2017)
  • 147 workers killed at work
  • 581,000 working people sustaining an injury at work according to the Labour Force Survey
  • 69,208 injuries to employees reported under RIDDOR
  • 28.2 million working days lost due to work-related illness and workplace injury
  • £15 billion estimated costs of injuries and ill health from current working conditions (2017/18)

( Source: https://www.hse.gov.uk/statistics/ )
Forklifts alone account for 1,300 UK employees being hospitalised each year (That’s 5 UK workers each workday!) with serious injuries due to these accidents. Unfortunately, that number is rising as there is a significant growth in e-commerce and warehouses across the UK. 

Use of AI-based computer vision for optimising workplace health and safety in the UK.
Our Computer Vision product, IRIS, is an AI-based computer vision solution to track and predict workplace accidents and then prevent them from happening. The existing use cases include:
  • Fork Lift Safety, Predicting and Prevention of Fork Lift accidents.
  • Use of Lifting Equipment
  • Work at Height
  • Fire & Thermal injuries and accidents
  • Machine Guarding
  • Manual Handling
  • Monitoring near misses and reporting near misses & accidents in real-time
  • Monitoring Use of PPE

IRIS is an enterprise computer vision solution. The product is currently deployed at many customer sites including Fortune 500 companies. The AI solution sits on the top of existing CCTV infrastructure. It is very cost-effective and can be deployed quickly either at a customer site or through the cloud. 

Monday, 22 June 2020

Covid-19 isn’t going away soon…… What’s your plan to make YOUR team safe in the workplace?

Thankfully Covid-19 seems to be on the wane…. But it will take a long time to disappear fully, and it is unlikely to be the last pandemic. This means that YOU may need to make the workplace safe for your team and perhaps your customers.

The good news is that many organisations have successfully implemented a Work-From-Home strategy….. And perhaps it has been much easier than we had dreamed it could be – certainly full or part-time WFH will now be a real option for many office workers.

The bad news is that solving the problem for workers who MUST be in their workplace is much, much harder.

Thankfully there ARE tools and solutions (automated Artificial Intelligence) that can help.

Solving the problem falls into several areas

  • Supplying / enforcing use of PPE
  • Providing hygiene solutions – hand sanitiser and washing facilities
  • Making social distancing easy & ensuring that it happens

The last of the above – social distancing, is easy to request, but much harder to implement & has a whole set of sub-problems

  • Filtering out people who have symptoms e.g. high temperature, that they themselves may be unaware of
  • Re-arranging the environment to provide enough isolation where individuals work
  • Reducing hotspots where people might struggle to keep a suitable distance
  • Changing processes to reduce face-face interactions where possible
  • Encouraging a culture where people choose to do the right thing

For these problems we have a technology solution: IRIS, an Artificial Intelligence tool that automatically and constantly analyses what is going on in your workplace AND gives you the data to work out what is going wrong, when it is going wrong and gives you the opportunity to fix it.

  • It can check who is & is not using PPE
  • It can check temperature of people entering a building
  • It can measure how far apart people are 24/7
  • It can find hotspots where social distancing guidelines being broken
  • Feedback is immediate and specific

Crucially this allows YOU to take control

  • Change processes exactly where it creates problems
  • Rearrange workspace to minimise squeeze points & hotspots
  • Make changes AND then check if these changes were effective
  • As with many processes you often get what you measure ….especially when the feedback loop is effective

Perhaps as importantly it allows you to SHOW that you are taking control, hopefully giving confidence to people, demonstrates your organisation’s desire for change and ideally creating a model for the right culture. 

Here is how you can find out more

Visit www.iwizardsolutions.com/covid19

Ask us at info.eu@iwizardsolutions.com

The Next Big thing: Post COVID-19 Workplaces

- Apoorva Verma

Pic: Freepik


The economy cannot survive a lockdown mode forever and the vaccine for COVID-19 may take years to develop and market, ‘unlock 1.0’ had become the necessary evil. This puts the responsibility on organizations, irrespective of their size, to focus on making the workplace safer for employees. 

 

This pandemic has changed the fundamental aspects of professional lives — from the daily commute to work environment to interactions amongst colleagues and peers. Hygiene and safety concerns are here to stay and will probably remain for at least a year from now or even more. 

 

Consequently, businesses are re-accessing their business models and working on best practices to follow on restarting operations. Thus, they are looking at:

 

1.    Ensuring compliance: what will be the best practices for floor management in a post-COVID-19 workplace? Face masks, sanitizer, gloves, and face shields are the new survival kits. Social distancing is being jacked up and automation solutions have seen a jump owing to the difficulty in ensuring these social distancing norms. 

 

2.    Hygiene: Can they deploy contactless technology wherever possible and emphasize on frequent sanitization in high-touch places?

 

3.    Accelerated technology adoption: Each sector has different requirements and use of advanced technologies such as AI and IIOT can help manage operations better. 

 

Returning back to work will require efforts and everyone needs to be vigilant. The world is at a stage where people can’t afford to be lackadaisical in their approach. Organisations are looking for long-term solutions to enforce safety compliance, especially in workplace environments where thousands work together. 

 

Computer vision technology solution trained for detecting compliance such as IRIS AI, by Integration Wizards, is ensuring face mask and social distancing compliance for organisations. It works on any existing CCTV network and identifies cases of social distancing or face mask violations and sends real-time alerts to the relevant authority. A dashboard provides detailed analytics for a given time period. In addition, using face recognition technology, their newly launched attendance app – LogMyFace, is a contactless attendance solution. 

 

Change is perpetual. But sometimes, it is the only option. As the world around us is changing, it has provided us with a new canvas. It is time to paint a new picture. 


Get in touch to discover the best compliance solution for your business. 

Monday, 13 January 2020

Looking Back: The best of 2019

“If you are not prepared to risk the usual, you will have to settle for the ordinary” – Jon Rohn

This year saw substantial growth in artificial intelligence and computer vision adaptation as well as the demand to keep up with the technology and innovation in businesses. These were the key drivers that kept us going, and also continue to motivate us for the upcoming decade.

We completed six years and successfully graduated from the start-up club to become an enterprise. From receiving ISO certification and Nasscom membership to getting recognised by Google amongst a handful MDM providers, it was a year to be proud of.

Our revenue has propelled eight times and we have on-boarded about twenty clients in the last six years. Working with our strategic partners Nvidia & Microsoft, we cracked the code on machine learning and many from our team up-skilled. We developed algorithms to control automated drones and upgraded for our client goals.

Our product is being used by 20,000 users across the globe and we actively participated in global conferences in the US, Spain and India. We grew not only in talent, but numbers too. A start-up that began with five now has about 65 members and a 7000sq. Ft of state of the art development centre in India.

Another feather to our cap came with the addition of seven Fortune 500 companies such as PostNord, Dover, Heineken, Xerox to name a few. Moreover, our AI-powered computer vision product, IRIS is now the synthetic intelligent eye for some corporate giants that include one of India’s largest solar plants, automobile manufacturer, chemical manufacturer, jewellery retail chain, construction corporate, heavy engineering, as well as FMCG. Last but not least, we also made some great improvements to our website – that’s right, we had a revamp of our website.

And we are not stopping here. We are striving to achieve more associations, more growth in 2020. 



#2020goals #yearendreview #sixyears #anniversary

Friday, 3 January 2020

Moving towards an AI-enabled future



McKinsey Global Institute claims that artificial intelligence is contributing to a transformation of society 10 times faster and at 300 times the scale, or roughly 3000 times the impact of the Industrial Revolution.

This is observed by the upsurge in artificial intelligence and computer vision adaptation as well as the demand to keep up with the technology and innovation in businesses in the last few years. According to a report by IDC, aggressive investments have been made in cognitive and AI solutions and in fact, global investments are expected to reach $57.6 billion by 2021.

Such investments are catalysed by the advent of modern computer vision and image processing techniques. AI-powered computer vision coupled with hardware-based accelerators have opened up the possibilities of analysing images in real-time to identify objects and activities. 

Since, a typical CCTV image is more than 100,000 Bytes, in this context, it might be quite apt to surmise that a picture is worth hundred-thousand words! 

At present, there are over 500 million CCTV cameras installed and the number is expected to rise to over a Billion by 2021. While these cameras cover everything from manufacturing, yards, warehouses, retail outlets to several parts of modern cities, so far they have been used retrospectively for monitoring and forensic analysis.

However, there is substantial growth in their usage in various verticals. For instance, retail outlets are getting equipped with the capability of knowing their customer demographics, dwell time and even emotions. Even the government is contemplating their use in smart city initiatives as they could prove beneficial if suspect activities are filtered from the live CCTV footages. Likewise, manufacturing premises bolster their safety parameters by ensuring any hazardous non-compliance is actively analysed and reported. 

In fact, stepping up occupational health & safety for people at all levels is the new benchmark that some of the companies are trying to work towards. If this becomes a norm, it could make a sustainable difference in global OSH challenges and promise a brighter future. 

Thus, an AI-enabled future lies in the best use of distributed vision technologies while delving into the deeper end of machine learning and deep learning to explore and understand the potential of these technologies better.

If you are looking to explore how computer vision technology can be useful in your enterprise, check out IRIS AI by Integration Wizards Solutions.

Tuesday, 5 November 2019

Computer Vision: Enhancing Industrial Safety with AI



by Apoorva Verma

The AI revolution is here.

As artificial intelligence increasingly gains prominence, several sub-domains such as computer vision, machine learning, deep learning, internet of things, and analytics are some of the technologies that have propelled growth.

Out of these, computer vision is one of those technologies that enable the interpretation and understanding of the visual world for machines. With the help of digital images and deep learning models, computers react to what they 'see' by identifying and classifying objects. In fact, accuracy rates in recognising and responding to visual inputs for the technology has risen from 50% to 99% over the past decade. This means that such solutions could become indispensable for a range of applications across industries.

However, our focus of discussion is the use of computer vision technology in manufacturing, which now have the necessary means to achieve automated safety compliance.

A Computer vision solution, such as the IRIS, developed by Integration Wizards, to work with an existing CCTV network, would serve as an advanced and effective replica of the human eyes, with the added ability to identify and classify different objects or situations, and react accordingly, such as in the form of alerts.

For instance, the AI-powered solution ensures workforce safety compliance by identifying workers without prerequisite safety equipment or protective gear such as hardhats, visibility vests, etc. This results in an appropriate response, like sending a real-time notification to the safety manager. The solution also maintains a database of safety protocol breaches which would be useful in investigations of any workplace accidents as well as take a step towards preventing accidents.

The application of the solution further extends from safety gear detection to occurrence of serious incidents such as the detection of fire and electrical malfunction, machine malfunction, trespassing or unauthorised access to hazardous areas, etc.

Efficient response trigged from the system in such scenarios aids in preventing serious losses to the workers as well as the manufacturing plant. Thus, it detects any anomalies that are not in accordance with the standard operating procedures. Real-time alerts and fail-safe measures accelerate the resolution of the issue.

The application of the solution can further encompass operational safety compliance. This would include material safety, such as the multi-object detection through computer vision, with automatic scanners on production lines, etc. In addition, it would identify any faults with raw materials that may be too small for the human eye but could prove detrimental for the final product.

In high performing manufacturing plants, compliance with safety regulations becomes the utmost priority. In fact, components falling off the production line must also adhere to safety guidelines.

Ultimately, the pressure of delivering high quality, efficient and time-sensitive results at manufacturing premises, together with the use of heavy machinery, potentially dangerous equipment, and the possibility of human error, make such sites prone to oversight of safety compliance, and by extension, workplace accidents.

Such unique innovative solutions can ensure safety compliance across the workforce as well as the entire manufacturing process and facility.