While our previous demands of our mobile cameras focused mainly on the number of megapixels and variety of filters, AI is now starting a true revolution in smartphone camera technology.
Thanks to AI, it is no longer necessary to have to enhance our photos after the fact. AI is now taking the role of photographer, making adjustments and predictions as the photos are taken in real time. This means that anyone with an advanced AI smartphone camera can become a pro photographer – no training, skills, or expensive equipment required.
The intuition of AI
AI-powered cameras like the Huawei P20 series intelligently assist users to take great photos by automatically optimising the camera settings on the backend. By learning about users’ behaviours and giving them what they want before they ask for it, the AI powering the devices is becoming increasingly sophisticated. With its enhanced object recognition capabilities, the AI smartphone can automatically determine the shooting modes and parameter settings for different scenes, and learn users’ photo-taking habits in order to deliver a more professional imaging effect. The more photos that are taken, the better the camera can interpret complicated environments.
The Huawei P20 and Huawei P20 Pro’s AI-powered real-time scene and object recognition technology can recognise more than 500 scenarios in 19 18 categories for users to choose the right skills at the right moment, such as the beach, food, or people. Once an object or a scene is identified, the smartphone automatically makes adjustments – from focus to contrast, to brightness, to colour – to produce the best possible results, regardless of the actual conditions you are presented with. The camera changes settings as it takes photos based on the information it is given; you do not have to browse through menus and select the settings you want.
For example, if a user is shooting at the beach, the AI-powered camera can identify objects and intelligently adjust for the best lighting, so the output looks good regardless of lighting conditions. Also, when the AI camera registers an attempt to take a group photo or a picture of the skyline, the AI Assisted Composition feature automatically draws a grid for the convenience of the user.
This AI-powered camera is capable of suggesting optimal parameters based on data gathered in the scene. Using machine learning, the devices also understand the photography habits of users over time, making suggestions based on various actions. AI also brings intelligent behaviour prediction, including intelligent context awareness and intelligent resource allocation.
4D Predictive Focus, breathtaking night shots and selfie evolution
One of the situations that has been the bane of photographers’ careers for decades now is night shots or low-light conditions. Because these scenarios usually require long exposure, these photos often come our out blurry, especially if they are taken with cellphones. AI image stabilisation technology, which is featured in the P20 devices, can eliminate these shaky shots and enable steady handheld night shots for incredible long exposure images. No tripod required!
4D predictive focus is another useful AI smartphone feature. In this mode, the camera predicts moving objects and focuses on them with extreme efficiency in order to capture minute details of objects in motion and eliminate blurred action shots. For the Huawei P20 and P20 Pro, the laser sensor provides instant focus with an effective range of three meters.
AI can also take the good old selfie to a whole new level with AI beautification. AI beautification provides 3D portrait lighting, making adjustments to better match the users’ facial features, resulting in luminous skin-tone enhancements and radiant selfies, even when backlit. These features are enhanced by the Huawei P20’s 24MP front camera, which captures the smallest details, even in low-light settings, bringing to life selfies with natural skin tone enhancements, delicately outlined face shapes, and 3D facial feature adjustments.
Other notable features of the Huawei P20 Pro include:
· The world’s first Leica triple camera, comprised of a 40MP RGB sensor, a 20MP monochrome sensor, and an 8MP sensor with telephoto lens. This results in images with more light, intricate detail, and greater zoom
· f/1.8, f/1.6 and f/2.4 wide aperture Leica lenses. The size of the aperture regulates how much light goes into the image sensor. The wider the aperture, the more light strikes the sensor, and therefore, the brighter and clearer the image – even in low-light environments.
· A brand new Leica 3x telephoto (VARIO-SUMMILUX-H 1:1.6-2.4/27-80ASPH) lens for long-range photography of up to 5x Hybrid Zoom.
· A revolutionary and highly-sensitive image sensor of 1/1.7 inches, 170 percent larger than the iPhone X, with a pixel size of 2 μm micrometers – 2.68 times higher than the iPhone X’s 1.22 μm
· Ultra Snapshot – devices can capture an image in as little as 0.3 seconds, even from a locked screen.
· Six-axis stabilisation and 960fps super slow motion, which is 4x higher than the standard 240fps and captures details that cannot be seen with the naked eye.
Smart home arrives in SA
The smart home is no longer a distant vision confined to advanced economies, writes ARTHUR GOLDSTUCK.
The smart home is a wonderful vision for controlling every aspect of one’s living environment via remote control, apps and sensors. But, because it is both complex and expensive, there has been little appetite for it in South Africa.
The two main routes for smart home installation are both fraught with peril – financial and technical.
The first is to call on a specialist installation company. Surprisingly, there are many in South Africa. Google “smart home” +”South Africa”, and thousands of results appear. The problem is that, because the industry is so new, few have built up solid track records and reputations. Costs vary wildly, few standards exist, and the cost of after-sales service will turn out to be more important than the upfront price.
The second route is to assemble the components of a smart home, and attempt self-installation. For the non-technical, this is often a non-starter. Not only does one need a fairly good knowledge of Wi-Fi configuration, but also a broad understanding of the Internet of Things (IoT) – the ability for devices to sense their environment, connect to each other, and share information.
The good news, though, is that it is getting easier and more cost effective all the time.
My first efforts in this direction started a few years ago with finding smart plugs on Amazon.com. These are power adaptors that turn regular sockets into “smart sockets” by adding Wi-Fi and an on-off switch, among other. A smart lightbulb was sourced from Gearbest in China. At the time, these were the cheapest and most basic elements for a starter smart home environment.
Via a smartphone app, the light could be switched on from the other side of the world. It sounds trivial and silly, but on such basic functions the future is slowly built.
Fast forward a year or two, and these components are available from hundreds of outlets, they have plummeted in cost, and the range of options is bewildering. That, of course, makes the quest even more bewildering. Who can be trusted for quality, fulfilment and after-sales support? Which products will be obsolete in the next year or two as technology advances even more rapidly?
These are some of the challenges that a leading South African technology distributor, Syntech, decided to address in adding smart home products to its portfolio. It selected LifeSmart, a global brand with proven expertise in both IoT and smart home products.
Equally significantly, LifeSmart combines IoT with artificial intelligence and machine learning, meaning that the devices “learn” the best ways of connecting, sharing and integrating new elements. Because they all fall under the same brand, they are designed to integrate with the LifeSmart app, which is available for Android and iOS phones, as well as Android TV.
Click here to read about how LifeSmart makes installing smart home devices easier.
Matrics must prepare for AI
By Vian Chinner, CEO and founder of Xineoh.
Many in the matric class of 2018 are currently weighing up their options for the future. With the country’s high unemployment rate casting a shadow on their opportunities, these future jobseekers have been encouraged to look into which skills are required by the market, tailoring their occupational training to align with demand and thereby improving their chances of finding a job, writes Vian Chinner – a South African innovator, data scientist and CEO of the machine learning company specialising in consumer behaviour prediction, Xineoh.
With rapid innovation and development in the field of artificial intelligence (AI), all careers – including high-demand professions like engineers, teachers and electricians – will look significantly different in the years to come.
Notably, the third wave of internet connectivity, whereby our physical world begins to merge with that of the internet, is upon us. This is evident in how widespread AI is being implemented across industries as well as in our homes with the use of automation solutions and bots like Siri, Google Assistant, Alexa and Microsoft’s Cortana. So much data is collected from the physical world every day and AI makes sense of it all.
Not only do new industries related to technology like AI open new career paths, such as those specialising in data science, but it will also modify those which already exist.
So, what should matriculants be considering when deciding what route to take?
For highly academic individuals, who are exceptionally strong in mathematics, data science is definitely the way to go. There is, and will continue to be, massive demand internationally as well as locally, with Element-AI noting that there are only between 0 and 100 data scientists in South Africa, with the true number being closer to 0.
In terms of getting a foot in the door to become a successful data scientist, practical experience, working with an AI-focused business, is essential. Students should consider getting an internship while they are studying or going straight into an internship, learning on the job and taking specialist online courses from institutions like Stanford University and MIT as they go.
This career path is, however, limited to the highly academic and mathematically gifted, but the technology is inevitably going to overlap with all other professions and so, those who are looking to begin their careers should take note of which skills will be in demand in future, versus which will be made redundant by AI.
In the next few years, technicians who are able to install and maintain new technology will be highly sought after. On the other hand, many entry level jobs will likely be taken care of by AI – from the slicing and dicing currently done by assistant chefs, to the laying of bricks by labourers in the building sector.
As a rule, students should be looking at the skills required for the job one step up from an entry level position and working towards developing these. Those training to be journalists, for instance, should work towards the skill level of an editor and a bookkeeping trainee, the role of financial consultant.
This also means that new workforce entrants should be prepared to walk into a more demanding role, with more responsibility, than perhaps previously anticipated and that the country’s education and training system should adapt to the shift in required skills.
The matric classes of 2018 have completed their schooling in the information age and we should be equipping them, and future generations, for the future market – AI is central to this.