Point set registration is one of the challenging tasks in areas such as pattern recognition, computer vision and image processing. Efficient performance of this task has been a hot topic of research due to its widespread applications. We propose a parameterised quantum circuit learning approach to point set matching problem. The proposed method benefits from a kernel-based quantum generative model that: 1) is able to find all possible optimal matching solution angles, 2) is potentially able to show quantum learning supremacy, and 3) benefits from kernel-embedding techniques and integral probability metrics for the definition of a powerful loss function. Moreover, the theoretical framework has been backed up by satisfactory preliminary and proof of concept experimental results.
翻译:点集登记是模式识别、计算机视觉和图像处理等领域的一项具有挑战性的任务。这项任务的高效执行由于其广泛应用而成为研究的一个热题。我们提出了一个参数化的量子电路学习方法,以找出匹配问题。拟议方法得益于基于内核的量子变现模型:(1) 能够找到所有可能的最佳匹配解决方案角度,(2) 有可能显示量子学习至上,(3) 内核组合技术和确定强力损失功能的综合概率衡量标准的好处。此外,理论框架还得到了概念实验结果令人满意的初步证据的支持。