The Burrows-Wheeler Transform (BWT) is a core technology in many modern compression and bioinformatics applications. Constructing the BWT for dynamically growing string collections remains a challenge. The existing optimal BWT (optBWT) construction algorithm for string collections can significantly reduce the number of BWT runs r for a given string collection. However, it requires reconstructing the entire BWT when new strings are added. To address this issue, this paper proposes an online BWT construction algorithm based on dynamic insertion interval — onptBWT (Online Computation of Near-Optimal BWT). This algorithm requires only O(r log n) bits of space and O(m log r) time (where n is the dataset length, m is the length of the newly added string, and r is the number of runs) for each new string added to produce a BWT that is near-optimal by only a small margin compared to optBWT in terms of runs r. Experimental results show that, across seven real-world genomic datasets, the average number of runs produced by onptBWT is only 1.41% higher than that of optBWT, outperforming six other BWT construction algorithms. In scenarios where the newly added strings are significantly smaller than the original string collection, onptBWT achieves faster construction speed and lower peak memory usage compared to the offline optBWT algorithm. Source code is available at https://github.com/xiaoYu0103/onptBWT.git.

A blockchain is a distributed ledger that allows users to exchange information without a centralized authority. This technology enables users to send and receive tokens among other applications, such as transactions, product management, and elections. It is possible to send data and tokens inside a single blockchain, but a method to efficiently share the data and tokens among different blockchains has not yet been constructed. Cross-chain communication, the focal point of several recent research efforts, is a scheme for sending data or tokens among different blockchains. In existing studies, a trusted third party (TTP) is used to ensure fair rates of token exchange among different blockchains. However, because blockchains are originally designed with a policy that does not incorporate the use of TTPs, the fair exchange rate should not be determined by TTPs, but rather by the market price of tokens among users. When exchange rates are determined from quotes among users, the preferred scheme is to determine the exchange rate offered by many users as an auction. Here, some existing cross-chain communication systems use smart contracts that automatically execute arbitrary processes on the blockchain. However, such schemes require a gas fee each time a smart contract is executed. Thus, implementing an auction scheme that determines the fair exchange rate among different blockchains would necessitate each user to pay a fee for each new token offered, which would result in high gas fees. In this study, we propose a scheme to determine exchange rates from quotes among users with a relatively low gas fee. Using a first-price sealed-bid auction and commitment scheme, the user with the highest token value can be identified without revealing the other users’ token offer values. In our scheme, the largest token value among users is determined as the exchange rate using an external Smart Contract (SC) instead of a TTP. We further modify the existing insert key-value commitment scheme to aggregate the commitment values of token offers. Our scheme is based on the generalized RSA assumption. By proving that it satisfies the key-binding property, we prove that the token sender cannot act maliciously. We further implement the proposed scheme and demonstrate that the gas fees and data space required to implement the proposed scheme are practically feasible.

In the past few years, research on lightweight block ciphers as security ciphers in the Internet of Things (IoT) has attracted considerable attention in cryptography. In this paper, we present an improved framework for neural distinguishers in lightweight SPN block ciphers suitable for IoT, focusing on two aspects: training data format and neural network structure. First, we analyze the nature of the SPN round function, then divide it into three cases to apply data augmentation. Second, we generate training data samples using three dimensions and construct neural networks using two-dimensional convolution. Finally, we validate the advantages of the improved framework on the SKINNY family and MIDORI family with higher accuracy and achieve a breakthrough in the number of rounds.

Precipitation is a crucial component of the natural water cycle, and inadequate timeliness and precision in precipitation prediction can result in agricultural losses, traffic disruptions, flood catastrophes, and even threats to human life. Consequently, precipitation prediction is a key problem in the domain of meteorology. However, the current methodologies pay close attention to the explicit spatial connections of precipitation regions while neglecting the implicit spatial connections over time. There are often challenging for traditional convolutional neural networks and graph neural networks to capture, leading to inaccurate spatial regions and poor timeliness of model predictions. To resolve this problem, we propose a Dynamic spatial-temporal graph prediction model for short-term precipitation (Dst-pred), which dynamically explores implicit connections among meteorological stations in the target region through graph neural networks and constructs dynamic spatial-temporal graphs to predict precipitation in the region. We have verified our Dst-pred model on our proprietary precipitation dataset from Guangxi Province, China, and the ERA5-Land dataset, and it can extract the implicit spatial connections between individual stations from the precipitation data of meteorological stations. The precipitation process capture of our model enhances the timeliness and accuracy of nowcasting precipitation prediction with the best performance.

We introduce innovative design concepts to address hard segmentation cases in complex image backgrounds and small training data sizes. We define (1) the shape feature description, including the medial surface, mask segmentation, and contour shape, which characterize articular cartilage shape, and establish a cartilage segmentation network to achieve multi-task consistency; and (2) the shape prior description, representing the shape distribution of articular cartilages, and establish a neural network based on this description. We incorporate the shape prior network into the multi-task consistency segmentation network. This results in a deep learning framework with high accuracy and strong generalization, guided by shape feature description and constrained by shape prior description. Our framework handles difficult cases with low contrast, ambiguous boundaries, deformed portions, and touching cartilages. The effectiveness of our method is demonstrated on two public knee image datasets and one clinical hip image dataset, where our approach shows increased segmentation accuracy compared to other state-of-the-art methods. Furthermore, its generalization is demonstrated for a subset of the BTCV dataset focusing on three specific structures: the aorta, the inferior vena cava, and the portal and splenic veins.

This letter presents a vertical-based adaptation of SPADE on Spark that significantly minimizes inter-worker communication. We achieve up to 6.2 × speedup over Spark MLlib’s PrefixSpan, enabling more efficient sequential pattern mining with minimal data movement and strong performance in distributed environments.

In this paper, we propose a method that combines an anisotropically weighted version of the directional relative total variation measure (AW-dRTV) with a pixel-wise Intensity and Scale Adjustable Edge-Preserving Smoothing filter (ISES filter). This approach enables effective removal of high-contrast textures with low computational cost while preserving important structural edges.

Title-overlaid images are useful as thumbnails on social media, where users prefer concise information to share and watch contents. Focusing on food contents, we aim to support creation of attractive title-overlaid food images to attract viewers’ attentions. This paper first analyzes the effect of font styles of the title on the attractiveness of title-overlaid images via preference experiments, and creates a dataset. Next, we design a prototype model of attractive font selection for a food image and its title. Its effectiveness is demonstrated through experiments on the created dataset.

This study presents design-rule compliant optical system guidelines for laser speckle authentication, clarifying key parameter constraints. We reveal the relationship between sampling and speckle size, and experimentally show enhanced robustness against displacement. Quantitive analysis using false acceptance and false rejectance rates confirms reliable performance of our design-rule compliant system.

Hebbian learning limits Hopfield network storage capacity (pattern-to-neuron ratio around 0.14). We propose Kernel Logistic Regression (KLR) learning. Unlike linear methods, KLR uses kernels to implicitly map patterns to high-dimensional feature space, enhancing separability. By learning dual variables, KLR dramatically improves storage capacity, achieving perfect recall even when pattern numbers exceed neuron numbers (up to ratio 1.5 shown), and enhances noise robustness. KLR demonstrably outperforms Hebbian and linear logistic regression approaches.