Protein structure prediction is a core challenge in computational biology, particularly for fragments within ligand-binding regions, where accurate modeling is still difficult. Quantum computing offers a novel first-principles modeling paradigm, but its application is currently limited by hardware constraints, high computational cost, and the lack of a standardized benchmarking dataset. In this work, we present QDockBank—the first large-scale protein fragment structure dataset generated entirely using utility-level quantum computers, specifically designed for protein–ligand docking tasks. QDockBank comprises 55 protein fragments extracted from ligand-binding pockets. The dataset was generated through tens of hours of execution on superconducting quantum processors, making it the first quantum-based protein structure dataset with a total computational cost exceeding one million USD. Experimental evaluations demonstrate that structures predicted by QDockBank outperform those predicted by AlphaFold2 and AlphaFold3 in terms of both RMSD and docking affinity scores. QDockBank serves as a new benchmark for evaluating quantum-based protein structure prediction.

The last decade has witnessed great advances in digital images. These images are subjected to many processing stages during storing, transmitting, or sharing over a network connection. Unfortunately, these processing stages could potentially add visual degradation to original image. These degradations reduce the perceived visual quality which leads to an unsatisfactory experience for human viewers. Therefore, image quality assessment (IQA) has become a topic of high interest and intense research over the last decade. This study mainly focuses on the most challenging category of IQA general‐purpose No‐Reference Image Quality Assessment (NR‐IQA), where the goal is to assess the quality of images without information about the reference images and without prior knowledge about the types of distortions in the tested image. A novel NR‐IQA approach is presented, by utilizing multiple deep belief networks (DBNs) with multiple regression models. It consists of four DBNs. Each DBN is associated with one type of distortion. The authors have evaluated the performance of the proposed and some existing models on a fair basis. The obtained results show that their model gives better results and yield a significant improvement.

This paper presents a computational performance analysis of an accelerated medical image registration using Graphics Processing Units (GPUs). In our previous work, a multi-resolution approach using normalized mutual information (NMI) has proven to be useful in medical image registration. In this paper, we propose an acceleration of the NMI procedure using GPU implementation because of the parallel processing capabilities. Registration algorithms were implemented on NVIDIA's GeForece 9600 GT graphic processor with the Compute Unified Device Architecture (CUDA) programming environment. Experimental results showed that the GPU implementation improves the registration computational performance with a speedup factor of 23.4x. In addition, the maximum speedup can be achieved with diligent data profiling.

This paper describes an effective framework to perform image segmentation and find regions of interest (ROI) in a user input object in an interactive way. Similar image objects are then retrieved from a repository. The repository stores off-line trained feature data of image objects, which was obtained by applying feature extraction and dimension reduction analysis to the ROI. The advantage of our framework is that the task is divided into two independent modules that can be implemented individually. Objects are processed in the image domain and the feature domain respectively. The framework has been successfully applied to image databases and can be easily modified to accommodate video objects.

This paper proposes a wavelet-domain hierarchical hidden Markov model for an unsupervised texture segmentation. Based on a hybrid graph structure, the global dependencies can be captured by a quad-tree structure across all scales, and local dependencies at higher resolution scales can be captured by a pyramidal graph structure. A novel context model that includes different positions, orientations, and scales is introduced. Applications of an unsupervised texture segmentation are presented. Compared with other alternative approaches for several test images, this method can achieve a significant improvement in segmentation, especially at higher resolution scales.

Wavelet domain hidden Markov models (WHMMs) provide a powerful approach for image modeling and processing because of the clustering and persistence properties of wavelet coefficients. However, the shift-variance of real wavelet transforms degrades the accuracy of the WHMMs. To overcome this problem, we propose a hidden Markov model based on the dual-tree complex wavelet transform that is approximately shift-invariant. Context information is used in this model to indicate the local correlation among wavelet coefficients. According to different visual attributes, several contexts based on frequency, orientation and scale are applied to capture both intrascale and interscale dependencies. The parameters of this model are estimated by an EM algorithm. Applications to image denoising are presented. The denoising performance is among the best state-of-the-art techniques, and outperforms models, which are based on real discrete wavelet transforms (DWTs).

A new texture image segmentation algorithm, HMTseg, was recently proposed and applied successfully to supervised segmentation. In this paper, we extend the HMTseg algorithm to unsupervised multiscale segmentation. A Gaussian mixture density is applied to approximate each wavelet coefficient's joint statistics through modeling each scale and each subband's histogram. A multiscale Expectation Maximization (EM) algorithm is used to integrate the parameter estimation and classification into one. Then, by the Hybrid Contextual Labeling Tree (HCL T), a Bayesian interscale and intrascale fusion algorithm is applied to raw segmentation results to obtain accurate and reliable final segmentations.

In this paper, a new repetition finder to be used with dynamic Huffman (1952) coding is proposed to improve the compression efficiency by reducing the redundancy due to string repetitions. Compared to the repetition finder proposed by Yokoo (1991), the proposed scheme effectively increases the numbers of consecutive symbols in the repetition mode and the total number of symbols in the repetition mode. Experimental results show that the proposed method outperforms the repetition finder of Yokoo by 14-40% in compression ratios with about the same memory requirement and running time.

A modified short-kernel filter pair is proposed for perfect reconstruction of HDTV signals. An interband prediction scheme based on the proposed filter pair is suggested to further reduce the average entropy of subband luminance signals. Simulations are conducted, and a modest reduction of the average entropy is obtained.>

Given a set of training samples, the commonly used approach to determine a universal model is accomplished by averaging the statistics over all training samples. It is suggested to use average divergence as a measurement for the effectiveness of a universal model and a minimax universal model that minimizes the maximum average divergence among all training samples is proposed. Efficient searching algorithms are developed and experimental results are presented.>

The encoding schemes utilize the first- and second-order Markov models to describe the source structure. Two coding techniques, Huffman encoding and arithmetic encoding, are used to achieve a high coding efficiency. Universal code tables which match the statistics of contour line drawings obtained from 64 contour maps are presented and can be applied to encode all contour line drawings with chain code representations. Experiments have shown about a 50% improvement on the code amount over the conventional chain encoding scheme with arithmetic coding schemes, and also have shown a compression rate comparable to that obtained by T. Kaneko and M. Okudaira (1985) with Huffman coding schemes, while this implementation is substantially simpler.>

A method for classifying closed planar shapes is presented. A shape is preprocessed and represented by several ordered sequences of vertices which are obtained by using an optimal polygon approximation algorithm with different approximation error criteria. A dynamic alignment algorithm is used to compute a similarity index between two sets of shape descriptors. The shape recognition process is hierarchical and invariant to rotation, translation and scaling. Classification experiments using noisy contours and objects have been performed with satisfactory results.>