Making Artificial Intelligence Accessible to ALL
We are privileged to live in an age where the data, compute and deep learning algorithms that are absolutely necessary to make AI a reality have all become abundant. The evolution of deep learning for AI has led to incredible advances in nearly all fields of technology and holds incredible promises for the future. Better healthcare, driverless cars, intelligent flying cars, improved access to financial resources and more informed allocation of state resources are just a few of what can be in the next decade.
But just like the early days of computer technology, the use of state-of-the-art AI is locked out of the reach of millions of developers. At present, even the most popular deep learning frameworks all require a great level of expertise when a developer needs to build object detection and recognition systems. Concrete knowledge of the foundations of deep learning algorithms is still a pre-requisite to building a sophisticated AI system. As a consequence, deep learning expertise requires a long period of intense study, and is often frustrating for most new comers who get lost in a mirage of concepts. “Confused” is the word many new entrants use to describe their situation has they attempt to go beyond simply training MNIST and CIFAR10 datasets for better accuracy.
In the same way, developing computer software initially required a deep understanding of the entire architecture of computers and operating systems. This frustrated early programmers including the legendary developers at Bell Labs. Realizing that the abstraction was tool low level, a number of experts created high level programming languages including Fortran, C, C++,etc. These languages were much easier to use than assembly languages but still required great expertise to use, hence, innovation was slow and computer programming was out of the reach of most people. Years later, high level programming languages like Python,Java,.NET, PHP, Javascript and more, made computer programming accessible to every single person and organization on the planet. Without these high level abstractions, the digital world as we know it today, will not exist.
Olafenwa Moses and I like to think of the current state of deep learning as the way programming was in the early days of C++.
If everyone can code, then everyone can build AI with the right tools. Hence, we both decided to dedicate ourselves to creating tools that will enable anyone, from average coders to expert industry professionals in various fields to integrate Artificial Intelligence into every solution they build.
Only when such tools exist, can we guarantee that AI will benefit everyone. We envision a Smart Future where every single application, device and system is infused with artificial intelligence. Access to artificial intelligence is a basic fundamental human right.
In light of this, we set out to build ImageAI, a very simple to use python computer vision library, that allows developers with absolutely no prior experience with ML or DL, to build state-of-the-art AI systems with just few lines of code.
Over the past three months since we released the first version, ImageAI has been used by thousands of developers all over the world, many of which experienced Artificial Intelligence for the first time. With ImageAI, any developer can perform Object detection, extraction and recognition in just 10 lines of python code!, any developer can train image recognition models with their own custom dataset in just 5 lines of python code!. Object Detection in Video in just 7 lines of code! All of these are backed with state-of-the-art RetinaNet, YoloV3, Resnet, Densenet, InceptionV3 and Squeezenet architectures.
To get a feel for how this works, I will actually go over some of these features.
First install ImageAI and its dependencies by following instructions from the official documentation.
First, Object Detection in 10 lines of code!
Download the yolo.h5 model file and put it in the same path as your .py file. Now you can run the detection on any image “image.jpg” in the above code.
The simple code above, will give you this:
Object Detection in Video in 7 lines of Code
This will give you this
ImageAI also supports many powerful features including advanced video analysis with interval callbacks. It fully supports object detection from IP Cameras and WebCams.
All of these amazing features are absolutely free and open source.
You can learn more from https://imageai.readthedocs.io or https://imageai-cn.readthedocs.io (Chinese version)
Our mission which we choose to accept is to advance, democratize and make artificial intelligence accessible to every single individual and corporate entities of all sizes on earth.
We are working on building more AI tools and bringing them to more platforms in the near future.
You can learn more about our mission on
Together, we can build a better world where Artificial Intelligence will continually transform our lives and be accessible to everyone.
You can reach to us on @johnolafenwa and @OlafenwaMoses
emails: guymodscientist@gmail.com and johnolafenwa@gmail.com
