Quantum computing is an emerging field of technology that utilizes unique properties such as the superposition principle and quantum entanglement, potentially enabling the development of techniques that are drastically faster or even the only feasible solutions compared to classical approaches. Thus, it promises enormous gains in several domains, including the processing of n-dimensional images, such as those produced by X-ray computed tomography. However, the latter is subject to physical effects that primarily include artifacts and quantum noise. Noise, in particular, is an unavoidable issue that needs to be addressed when processing X-ray data. Therefore, this work considers the denoising of 3D volumetric data using quantum computing. An adaptation of a mathematical denoising model was transformed to be suitable for quantum implementation, and preliminary experiments demonstrated its functionality with two-dimensional images. Here, we first lift this model to be applicable to 3D images. Next, we provide an implementation in form of quantum circuits and consider issues of a practical implementation occurring specifically in 3D. The final implementation is executed on a real X-ray computed tomography dataset, showing that proper denoising can be performed on quantum devices, yet current technological limitations inhibit the application to large datasets.
Quantum computing is an emerging field of technology that utilizes unique properties such as the superposition principle and quantum entanglement, potentially enabling the development of techniques that are drastically faster or even the only feasible solutions compared to classical approaches. Thus, it promises enormous gains in several domains, including the processing of n-dimensional images, such as those produced by X-ray computed tomography. However, the latter is subject to physical effects that primarily include artifacts and quantum noise. Noise, in particular, is an unavoidable issue that needs to be addressed when processing X-ray data. Therefore, this work considers the denoising of 3D volumetric data using quantum computing. An adaptation of a mathematical denoising model was transformed to be suitable for quantum implementation, and preliminary experiments demonstrated its functionality with two-dimensional images. Here, we first lift this model to be applicable to 3D images. Next, we provide an implementation in form of quantum circuits and consider issues of a practical implementation occurring specifically in 3D. The final implementation is executed on a real X-ray computed tomography dataset, showing that proper denoising can be performed on quantum devices, yet current technological limitations inhibit the application to large datasets.
Quantum computing is currently one of the fastest emerging branches of information processing due to its theoretical computation powers exceeding conventional computers by far. However, currently well-known quantum-powered algorithms are of theoretical nature and its effect on practical problems has yet to be discovered. This work reviews current approaches in context of the applicability of quantum computing for the domain of image processing with respect to 3D computed tomography. It focuses on the encoding of n-dimensional images on a quantum computer and also the quantum measurement process, i.e., the image read-out. First results show that while the encoding is indeed highly efficient and is well-suited for representing 3D voxel data as obtained from computed tomography, the decoding to a classical representation is rather expensive. The latter part is also very sensitive to quantum hardware noise as was evidenced by both noisy simulator and real quantum hardware experiments.
Positron emission tomography (PET) with long axial field-of-view (LAFOV) scanners generates unprecedentedly large datasets, posing major challenges for conventional reconstruction algorithms. Recent advances in deep learning have improved PET image quality, yet scaling to high-dimensional data remains computationally demanding. Quantum computing (QC) offers new opportunities to address such challenges. This work presents a pilot study on image denoising using a hybrid quantum-classical autoencoder. The model employs a classical encoder, a quantum bottleneck implemented via a parameterized quantum circuit, and a classical decoder. Using simulated experiments on the medical modified national institute of standards and technology (MedMNIST) dataset with 28 × 28 grayscale images and on PET scan images with 64 × 64, results demonstrated that the quantum bottleneck can effectively recover structural details from noisy inputs. While scaling to large-volume PET data will require advances in quantum hardware, circuit design, and error mitigation, these results highlight the potential of hybrid quantum–classical approaches to complement deep learning in PET reconstruction.
Quantum Image Processing (QIP) is a field that aims to utilize the benefits of quantum computing for manipulating and analyzing images. However, QIP faces two challenges: the limitation of qubits and the presence of noise in a quantum machine. In this research we propose a novel approach to address the issue of noise in QIP. By training and employing a machine learning model that identifies and corrects the noise in quantum processed images, we can compensate for the noisiness caused by the machine and retrieve a processing result similar to that performed by a classical computer with higher efficiency. The model is trained by learning a dataset consisting of both existing processed images and quantum processed images from open access datasets. This model will be capable of providing us with the confidence level for each pixel and its potential original value. To assess the model's accuracy in compensating for loss and decoherence in QIP, we evaluate it using three metrics: Peak Signal to Noise Ratio (PSNR), Structural Similarity Index (SSIM), and Mean Opinion Score (MOS). Additionally, we discuss the applicability of our model across domains well as its cost effectiveness compared to alternative methods.
Quantum-inspired computing combines conventional computational techniques with mathematical concepts motivated by quantum mechanics, such as superposition, entanglement-inspired correlations, and wave localization. This review examines recent developments in quantum-inspired methods for medical image denoising, edge detection, segmentation, feature extraction, secure image processing, and multimodal image fusion involving MRI, CT, and ultrasound images. Reported studies indicate that these methods can improve selected measures of image quality, segmentation, feature extraction, and classification under experimental settings. But, the variation of datasets, frameworks for evaluation of methods, baseline algorithms, and dimensions of data complicate the issue of comparing different methods, while low level of external validation does not allow one to conclude whether quantum methods have some advantages with respect to traditional classical and deep learning models.Most of the quantum-influenced methods are implemented on conventional computing systems that give a possibility to use the methods developed using classical image processing methods, and new learning methods, such as few-shot learning and domain adaptation. The current review provides discussion concerning new architectures and evaluation strategies in clinical practice and also potential problems with scalability, readiness of hardware and integration of work processes. One of the main problems is that many studies are conducted on synthetic datasets, and this makes it difficult to estimate the clinical relevance of the methods under consideration. Overall, this review presents a balanced view of the opportunities, current limitations, and requirements for translating quantum-inspired medical image processing into clinical practice.
Quantum Image Processing (QIP) is a field that aims to utilize the benefits of quantum computing for manipulating and analyzing images. However, QIP faces two challenges: the limitation of qubits and the presence of noise in a quantum machine. In this research, we propose a novel approach to address the issue of noise in QIP. By training and employing a machine learning model that identifies and corrects the noise in quantum-processed images, we can compensate for the noisiness caused by the machine and retrieve a processing result similar to that performed by a classical computer with higher efficiency. The model is trained by learning a dataset consisting of both existing processed images and quantum-processed images from open-access datasets. This model will be capable of providing us with the confidence level for each pixel and its potential original value. To assess the model's accuracy in compensating for loss and decoherence in QIP, we evaluate it using three metrics: Peak Signal to Noise Ratio (PSNR), Structural Similarity Index (SSIM), and Mean Opinion Score (MOS). Additionally, we discuss the applicability of our model across domains well as its cost effectiveness compared to alternative methods.
One of the primary difficulties in computed tomography (CT) is reconstructing cross-sectional images from measured projections of a physical object. There exist several classical methods for this task of generating a digital representation of the object, including filtered backprojection or simultaneous algebraic reconstruction technique. Our research aims to explore the potential of quantum computing in the field of industrial X-ray transmission tomography. Specifically, this work focuses on the application of a method similar to that proposed by Nau et al. (2023) on real CT data to demonstrate the feasibility of quadratic-unconstrained-binary-optimization-based tomographic reconstruction. Starting with simulated phantoms, results with simulated annealing as well as real annealing hardware are shown, leading to the application on measured cone-beam CT data. The results demonstrate that tomographic reconstruction using quantum annealing is feasible for both simulated and real-world applications. Yet, current limitations—involving the maximum processable size and bit depth of voxel values of the images, both correlated with the number of densely connected qubits within the annealing hardware—imply the need of future research to further improve the results. This approach, despite its early stage, has the potential to enable more sophisticated reconstructions, providing an alternative to traditional classical methods.