With the promising advancement in quantum computing technology in the last decade, there is a strong motivation to find suitable applications for quantum algorithms and quantum computers. Domains such as High Energy Physics (HEP) have an enormous readout count of high-resolution data. Performing pattern recognition on this readout is computationally challenging and time-consuming because of the multi-dimensionality of the data. In this paper, we propose a methodology that employs quantum algorithms such as Quantum Wavelet Transform and Grover's search algorithm for timeefficient pattern recognition in data sets that are characterized by high spatial resolution and high dimensionality. The motivation behind using quantum algorithms is the potential speedup relative to classical methods, when performed by a quantum computer. In our proposed methodology, Quantum Wavelet Transform is performed on the high spatial resolution data to reduce its dimensionality while quantum Grover's search algorithm is employed to search for target patterns in the reduced data set. Performing the search operation on data with reduced spatial resolution, minimizes processing overheads and computation times. Moreover, use of quantum techniques yield faster results, compared to classical dimension reduction and search methods. We demonstrate the feasibility of the proposed methodology by emulating the quantum algorithms on classical hardware based on field programmable gate arrays (FPGAs). A high performance reconfigurable computer (HPRC) was used for the experimental evaluation. The obtained results are favorable towards our proposed approach.
Quantum search algorithm proposed by Grover(Grover's algorithm) finds target data with equal probability for all target data. Y. Seo et al. [IEEE Access, vol. 10, pp.16209-16224 (2022)] proposed the quantum search algorithm for weighted solutions to discriminate the importance of target data. In the simple case, we represent the analysis of the algorithm. The low bound of the number of applying Grover operator
tis described using amplitude expression. If we choose an appropriate
tvalue, we provide the maximum ratio of the amplitude of high and low-importance target data. In addition, we arrange the maximum probability of high-importance target data using the bound.