Two companies based in Spain, Multiverse Computing and IKERLAN have partnered and released a paper describing how quantum computer vision can be used to detect defects in manufactured car pieces in a manufacturing production line. This would represent one of the first implementation of quantum computer vision for a relevant problem in a manufacturing production line. Although this research only created proofs-of concepts, it does point the way for utilization of quantum computing in manufacturing and quality control.

In a preprint titled, Quantum artificial vision for defect detection in manufacturing post on arXiv, the companies described two different approaches they have looked at for thsi problem. Both approaches look at data from individual parts and creates an algorithm that classifies each part as either defective or non-defective. The first is an algorithm called a quantum Support Vector Machine (SVM) intended for use on a gate-based processor. Due to long potential running times, the researchers tested this algorithm on the Qiskit State Vector Simulator. The second approach uses an algorithm called QBoost which runs on a D-Wave quantum annealer. This second approach has an added benefit in that model training can be run on the quantum computer, but then using the model after it has been trained can be performed on a classical computer. The researchers have concluded that both methods have the potential to outperform classical approaches in the identification of defects.

Additional details about this research can be seen in a news release available on the Multiverse website here and the full technical paper posted on arXiv here.

August 16, 2022