Quantum computing vision meets manufacturing production lines
Computer vision is everywhere in our lives. You can find it in automated passport scanners at airports, security systems for your smartphone, and even in parking lots to read your license plate. Computers are constantly watching us, and that’s not a bad thing. This type of “Big Brother” system allows computerized systems to make decisions quickly and efficiently, which saves a lot of time and human effort in complex tasks.
Computer vision is also relevant for a manufacturing production line for the same reasons. Imagine a factory that produces wheels for cars. Wheel quality assessment can be done manually by an employee, but it can also be done by computerized machine vision systems. Such systems are able to detect the smallest fractures and irregularities of the wheel. Computer vision is a key part of production lines in factories today.
Although machine vision offers some advantages to improve production, the current technology is not sophisticated enough for advanced manufacturing lines. One of the challenges is poor image quality due to older camera hardware, poor lighting, vibration, and factors found in most manufacturing facilities. Additionally, a machine vision system typically requires a very high level of accuracy to detect faults, as well as the ability to explain the results. Imagine going to tell your boss that you don’t know why a machine rejected a part!
Multiverse Computing, a provider of value-based quantum computing solutions, conducted a pilot project with manufacturing company Ikerlan evaluate the performance of quantum computer vision in detecting part defects in factory production lines.
Quantum computers are the holy grail of algorithmic science: they promise faster, more efficient and more accurate calculations compared to today’s conventional computers, because they manipulate information according to the laws of quantum physics. Quantum computing algorithms can detect anomalies in subtle patterns of data that a conventional computer would never be able to find. Quantum computers are also capable of extracting more information from less data, as evidenced by a recently published paper in Nature Communications.
The Multiverse Computing and Ikerlan teams demonstrated that today’s quantum computers, despite their hardware limitations, outperform conventional computer vision systems at finding defects in image fabrication. The results, published in a research paper on arXiv, show that quantum computer vision outperforms its classical counterparts in accuracy, even while maintaining algorithm training and inference times. From a broader perspective, this is one of the first breakthrough results showing that quantum computers can offer real commercial value to industry today for some specific applications.
Prove the power of quantum computing
Multiverse and Ikerlan used a dataset of 2,727 x-ray images of automotive parts with and without molding defects to train a quantum machine vision system to see whether an image has a defect or not. This is an incredibly difficult task for computer vision systems due to the low quality of some images and the complexity of the defects, but it is the necessary first step in identifying defects like a small fracture in a wheel. car which could possibly lead to an accident. The study concluded that some of today’s commercial quantum computers can run quantum classification algorithms that consistently detect these defects with higher accuracy (between 10% and 20% more) than traditional explainable machine learning methods, and with a similar inference time. The multiverse algorithms were also explainable, in that a human could understand the reasons for the decisions made by the algorithm. It is a must in many industrial applications.
In addition to outperforming classical computers, these quantum algorithms from Multiverse have learned to detect defects faster than other advanced algorithms currently being tested.
Compared to Amazon Solutions Lab’s best deep learning algorithms, the researchers found that the Multiverse/Ikerlan quantum algorithm achieved essentially the same accuracy but was 24 times faster to train, with the time needed dropping from nearly two hours. and a half hours to less than six minutes. This speedup is critical in environments where the algorithm needs to adapt to different circumstances on the fly. So overall, not only are quantum vision systems explainable, but they are also more accurate than their classical counterparts and significantly faster than deep learning.
The results of the Multiverse and Ikerlan collaboration are the first example of a quantum computer vision system applied to a manufacturing line. These results prove that quantum computing has quality control value in today’s factories. This is just the tip of the iceberg: the next time you go through automated passport control, there may be quantum physics behind the scan!