Refinement Strategy Research Articles

Superpixel-based multi-scale multi-instance learning for hyperspectral image classification

Superpixels can define meaningful local regions within a hyperspectral image (HSI) and have become the building blocks of various HSI classification methods. The superpixels in HSIs consist of multiple spectral pixels, sharing a similar structure with the data in multi-instance learning (MIL). However, the potential of MIL methods in the field of HSI classification has been rarely explored. In this paper, we propose the superpixel-based multi-scale multi-instance learning (MSMIL) framework, enhancing the superpixel representation with MIL for the first time. Segmenting the HSIs with superpixels, MSMIL converts the HSI classification into MIL problems and extracts superpixel representations via the MIL method, namely multi-instance factor analysis (MIFA). Compared with the existing methods focusing exclusively on the local information, MIFA utilizes the deviations from an overall generative model to describe the superpixels, retaining both the local and the global information. Moreover, MSMIL introduces multi-scale superpixels and a spectral–spatial decision fusion strategy for further refinement, where the results of multi-scale superpixel maps are weighted according to prediction certainty and spatial consistency. The proposed method is evaluated on four benchmark datasets and achieves competitive results. For example, MSMIL outperforms the comparison methods with a margin of 5% overall accuracy on the Indian Pines dataset, when 2% pixels are selected as the training set.

Improving the approximation quality of tensor product B-spline surfaces by local parameterization

Abstract Freeform surfaces like tensor product B-spline surfaces have been proven to be a suitable tool to model laser scanner point clouds, especially those representing artificial objects. However, when it comes to the modelling of point clouds representing natural surfaces with a lot of local structures, tensor product B-spline surfaces reach their limits. Refinement strategies are usually used as an alternative, but their functional description is no longer nearly as compact as that of classical tensor product B-spline surfaces, making subsequent analysis steps considerably more cumbersome. In this publication, the approximation quality of classical tensor product B-spline surfaces is improved by means of local parameterization. By using base surfaces with a local character, relevant information about local structures of the surface to be estimated are stored in the surface parameters during the parameterization step. As a consequence, the resulting tensor product B-spline surface is able to represent these structures even with only a small number of control points. The developed locally parameterized B-spline surfaces are used to model four data sets with different characteristics. The results reveal a clear improvement compared to the classical tensor product B-spline surfaces in terms of correctness, goodness-of-fit and stability.

Hybrid Intelligence for Marine Biodiversity: Integrating Citizen Science with AI for Enhanced Intertidal Conservation Efforts at Cape Santiago, Taiwan

Marine biodiversity underpins the formation of marine protected areas (MPAs), necessitating detailed surveys to account for the dynamic temporal and spatial distribution of species influenced by tidal patterns and microhabitats. The reef rock intertidal zones adjacent to urban centers, such as Taiwan’s Cape Santiago, exhibit significant biodiversity, yet they are increasingly threatened by tourism-related activities. This study introduces an artificial intelligence (AI)-empowered citizen science (CS) approach within the local community to address these challenges. By integrating CS with AI, we establish a hybrid intelligence (HI) system that conducts in situ biological surveys and educational programs focused on reef ecological conservation. This initiative not only facilitates the collective gathering and AI-assisted analysis of critical data but also uses machine-learning outputs to gauge data quality, thus informing subsequent data collection and refinement strategies. The resulting collectivity and iterative enhancement foster a mutual and continuous HI learning environment. Our HI model proves instrumental in fostering community engagement and public involvement in CS endeavors, cultivating the skills necessary for documenting rocky intertidal biodiversity shifts. These efforts are pivotal for informing the design and governance of future MPAs, ensuring their efficacy and sustainability in marine conservation.

Long-range proton and hydroxide ion transfer dynamics at the water/CeO2 interface in the nanosecond regime: reactive molecular dynamics simulations and kinetic analysis.

The structural properties, dynamical behaviors, and ion transport phenomena at the interface between water and cerium oxide are investigated by reactive molecular dynamics (MD) simulations employing neural network potentials (NNPs). The NNPs are trained to reproduce density functional theory (DFT) results, and DFT-based MD (DFT-MD) simulations with enhanced sampling techniques and refinement schemes are employed to efficiently and systematically acquire training data that include diverse hydrogen-bonding configurations caused by proton hopping events. The water interfaces with two low-index surfaces of (111) and (110) are explored with these NNPs, and the structure and long-range proton and hydroxide ion transfer dynamics are examined with unprecedented system sizes and long simulation times. Various types of proton hopping events at the interface are categorized and analyzed in detail. Furthermore, in order to decipher the proton and hydroxide ion transport phenomena along the surface, a counting analysis based on the semi-Markov process is formulated and applied to the MD trajectories to obtain reaction rates by considering the transport as stochastic jump processes. Through this model, the coupling between hopping events, vibrational motions, and hydrogen bond networks at the interface are quantitatively examined, and the high activity and ion transport phenomena at the water/CeO2 interface are unequivocally revealed in the nanosecond regime.

Enhancing word sense disambiguation through contextual embedding and optimization techniques���

This study presents a new strategy for determining the meaning of words based on their use in text. Traditional methods struggle because they do not fully grasp the context around ambiguous words. Our framework combines powerful models that understand context, like BERT and ELMO, to precisely identify a word’s sense based on its language environment. These representations depict word meanings in different situations more accurately. This helps the model to predict various meanings of an unclear word. Additionally, we apply techniques to refine the model’s performance. For instance, we customize the representations specifically for making sense of words. This increases the model’s awareness of subtle clues. Experiments show our approach notably outperforms basic methods in correctly determining word senses. The blending of contextual representations and refinement strategies not only boosts overall accuracy for defining ambiguous words but also demonstrates flexibility in handling difficult examples across diverse language contexts.

Fine-grained video super-resolution via spatial-temporal learning and image detail enhancement

This paper addresses the problem for fine-grained video super-resolution (FGVSR) to suppress temporal flickering caused by separately processed consecutive frames and enhance the quality of restored video frame details when upsizing videos. Some existing video SR methods fail to sufficiently utilize spatial-temporal information from input low-resolution (LR) videos, while others may generate undesirable artifacts or cannot well reconstruct image details. To overcome these problems, we present a novel deep learning framework for FGVSR, which takes a set of consecutive LR video frames and generate the corresponding super-resolved frames. Our deep FGVSR framework focuses on reconstructing missing information from the LR sources based on the proposed multi-frame alignment and refinement strategies. More specifically, we propose an alignment module, where multiple frames are aligned at feature level, to prevent the output videos from flickering. Then, we introduce a feature fusion module, where aligned features generated from our alignment module are fused and refined in a multi-scale manner. Finally, the proposed refinement module is used to reconstruct missing information based on the fused features. In addition, we also embed an image enhancement module on the skip connection from the input layer to the output layer of our network for further enhancing the SR results. Experimental results show that the proposed deep FGVSR, compared with existing deep learning-based VSR methods, achieves state-of-the-art performances on the three well-known benchmarks, including REDS, Vid4, and Vimeo90k. More specifically, compared with the state-of-the-art VSR methods in our experiments, our FGVSR achieves quantitative improvements from 0.70 dB to 9.54 dB in PSNR. On the other hand, our method has also been shown to be efficient to other image restoration tasks, such as image inpainting.

Enhancing Armenian Automatic Speech Recognition Performance: A Comprehensive Strategy for Speed, Accuracy, and Linguistic Refinement

This research introduces a comprehensive strategy to enhance the performance of an existing automatic speech recognition (ASR) model, which has been previously documented in published articles. The study sets out to achieve several objectives. Firstly, it concentrates on updating the ASR model by retraining it with new datasets. This involves integrating samples from the latest Common Voice corpus release and data collected independently via the armspeech.com web application. Another key focus lies in optimizing the ASR model for near-real-time processing, intending to improve its speed and efficiency. The proposed adjustments to the model’s architecture aim to balance accuracy and processing speed, which is essential for applications requiring prompt speech recognition. Furthermore, the research explores the integration of Transformer models into the post-processing pipeline to introduce punctuation and capitalization into the ASR output. This step not only enhances the linguistic quality of transcriptions but also improves their readability and usability. In tandem with these advancements, the research presents a systematic approach to gathering, annotating, and storing datasets specifically tailored for punctuation and capitalization tasks. The methodology outlines the acquisition and organization of a dataset conducive to training Transformer models for these linguistic tasks. This comprehensive approach, which encompasses dataset enrichment, architectural modifications, and post-processing enhancements, aims to elevate the ASR model’s accuracy, speed, and linguistic refinement, with a particular focus on addressing the intricacies of the Armenian language. The research contributes valuable insights into the optimization of ASR systems, tackling both language-specific challenges and broader issues related to linguistic post-processing.

Progressive Frame-Proposal Mining for Weakly Supervised Video Object Detection.

In this paper, we focus on the weakly supervised video object detection problem, where each training video is only tagged with object labels, without any bounding box annotations of objects. To effectively train object detectors from such weakly-annotated videos, we propose a Progressive Frame-Proposal Mining (PFPM) framework by exploiting discriminative proposals in a coarse-to-fine manner. First, we design a flexible Multi-Level Selection (MLS) scheme, with explicit guidance of video tags. By selecting object-relevant frames and mining important proposals from these frames, the proposed MLS can effectively reduce frame redundancy as well as improve proposal effectiveness to boost weakly-supervised detectors. Moreover, we develop a novel Holistic-View Refinement (HVR) scheme, which can globally evaluate importance of proposals among frames, and thus correctly refine pseudo ground truth boxes for training video detectors in a self-supervised manner. Finally, we evaluate the proposed PFPM on a large-scale benchmark for video object detection, on ImageNet VID, under the setting of weak annotations. The experimental results demonstrate that our PFPM significantly outperforms the state-of-the-art weakly-supervised detectors.

Adaptive solution of the domain decomposition+L2-jumps method applied to the neutron diffusion equation on structured meshes

At the core scale, neutron deterministic calculations are usually based on the neutron diffusion equation. Classically, this equation can be recast in a mixed variational form, and then discretized by using the Raviart-Thomas-Nédélec Finite Element. The goal is to extend the Adaptive Mesh Refinement (AMR) strategy previously proposed in [1] to the Domain Decomposition+L2 jumps which allows non conformity at the interface between subdomains. We are able to refine each subdomain independently, which eventually leads to a more optimal refinement. We numerically investigate the improvements made to the AMR strategy.

Robust Coarse Cage Construction With Small Approximation Errors.

We propose a robust and automatic method to construct manifold cages for 3D triangular meshes. The cage contains hundreds of triangles to tightly enclose the input mesh without self-intersections. To generate such cages, our algorithm consists of two phases: (1) construct manifold cages satisfying the tightness, enclosing, and intersection-free requirements and (2) reduce mesh complexities and approximation errors without violating the enclosing and intersection-free requirements. To theoretically make the first stage have those properties, we combine the conformal tetrahedral meshing and tetrahedral mesh subdivision. The second step is a constrained remeshing process using explicit checks to ensure that the enclosing and intersection-free constraints are always satisfied. Both phases use a hybrid coordinate representation, i.e., rational numbers and floating point numbers, combined with exact arithmetic and floating point filtering techniques to guarantee the robustness of geometric predicates with a favorable speed. We extensively test our method on a data set of over 8500 models, demonstrating robustness and performance. Compared to other state-of-the-art methods, our method possesses much stronger robustness.

Mutual Information Regularization for Weakly-Supervised RGB-D Salient Object Detection

In this paper, we present a weakly-supervised RGB-D salient object detection model via scribble supervision. Specifically, as a multimodal learning task, we focus on effective multimodal representation learning via inter-modal mutual information regularization. In particular, following the principle of disentangled representation learning, we introduce a mutual information upper bound with a mutual information minimization regularizer to encourage the disentangled representation of each modality for salient object detection. Based on our multimodal representation learning framework, we introduce an asymmetric feature extractor for our multimodal data, which is proven more effective than the conventional symmetric backbone setting. We also introduce multimodal variational auto-encoder as stochastic prediction refinement techniques, which takes pseudo labels from the first training stage as supervision and generates refined prediction. Experimental results on benchmark RGB-D salient object detection datasets verify both effectiveness of our explicit multimodal disentangled representation learning method and the stochastic prediction refinement strategy, achieving comparable performance with the state-of-the-art fully <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">supervised</i> models. Our code and data are available at: https://npucvr.github.io/MIRV/.

Vote2Cap-DETR++: Decoupling Localization and Describing for End-to-End 3D Dense Captioning.

3D dense captioning requires a model to translate its understanding of an input 3D scene into several captions associated with different object regions. Existing methods adopt a sophisticated "detect-then-describe" pipeline, which builds explicit relation modules upon a 3D detector with numerous hand-crafted components. While these methods have achieved initial success, the cascade pipeline tends to accumulate errors because of duplicated and inaccurate box estimations and messy 3D scenes. In this paper, we first propose Vote2Cap-DETR, a simple-yet-effective transformer framework that decouples the decoding process of caption generation and object localization through parallel decoding. Moreover, we argue that object localization and description generation require different levels of scene understanding, which could be challenging for a shared set of queries to capture. To this end, we propose an advanced version, Vote2Cap-DETR++, which decouples the queries into localization and caption queries to capture task-specific features. Additionally, we introduce the iterative spatial refinement strategy to vote queries for faster convergence and better localization performance. We also insert additional spatial information to the caption head for more accurate descriptions. Without bells and whistles, extensive experiments on two commonly used datasets, ScanRefer and Nr3D, demonstrate Vote2Cap-DETR and Vote2Cap-DETR++ surpass conventional "detect-then-describe" methods by a large margin.

Pathological Asymmetry-Guided Progressive Learning for Acute Ischemic Stroke Infarct Segmentation.

Quantitative infarct estimation is crucial for diagnosis, treatment and prognosis in acute ischemic stroke (AIS) patients. As the early changes of ischemic tissue are subtle and easily confounded by normal brain tissue, it remains a very challenging task. However, existing methods often ignore or confuse the contribution of different types of anatomical asymmetry caused by intrinsic and pathological changes to segmentation. Further, inefficient domain knowledge utilization leads to mis-segmentation for AIS infarcts. Inspired by this idea, we propose a pathological asymmetry-guided progressive learning (PAPL) method for AIS infarct segmentation. PAPL mimics the step-by-step learning patterns observed in humans, including three progressive stages: knowledge preparation stage, formal learning stage, and examination improvement stage. First, knowledge preparation stage accumulates the preparatory domain knowledge of the infarct segmentation task, helping to learn domain-specific knowledge representations to enhance the discriminative ability for pathological asymmetries by constructed contrastive learning task. Then, formal learning stage efficiently performs end-to-end training guided by learned knowledge representations, in which the designed feature compensation module (FCM) can leverage the anatomy similarity between adjacent slices from the volumetric medical image to help aggregate rich anatomical context information. Finally, examination improvement stage encourages improving the infarct prediction from the previous stage, where the proposed perception refinement strategy (RPRS) further exploits the bilateral difference comparison to correct the mis-segmentation infarct regions by adaptively regional shrink and expansion. Extensive experiments on public and in-house NCCT datasets demonstrated the superiority of the proposed PAPL, which is promising to help better stroke evaluation and treatment.

3D Modelling: A Review

Three-dimensional (3D) modeling is a transformative process in the digital realm that involves the creation of virtual representations of objects, environments, or characters. This multifaceted discipline plays a pivotal role across diverse industries, encompassing animation, gaming, virtual reality, architecture, product design, and more. The fundamental components of 3D modeling include vertices, edges, and faces, forming meshes that define the object's geometry. Techniques such as polygonal modeling, NURBS modeling, and subdivision surfaces offer varying approaches to sculpting digital forms. UV mapping ensures precise texture application, while rigging and animation breathe life into models through skeletal structures and movement. The integration of textures, materials, and lighting enhances the realism of rendered images or animations. Computer-Aided Design (CAD) modeling caters to engineering precision, and sculpting tools enable intricate detailing for characters or organic shapes. As technology advances, 3D modeling continues to evolve, influencing the way we visualize, design, and interact with virtual worlds. This abstract provides an overview of key concepts in 3D modeling, highlighting its significance in shaping digital experiences and pushing the boundaries of creative expression and technical innovation.

Adaptive meshfree method for fourth-order phase-field model of fracture using consistent integration schemes

An adaptive fourth-order phase-field model is proposed using consistent element-free Galerkin method (EFG). Convenient construction of high-order interpolation functions in EFG method is fully utilized. Requirement of C1 continuity of shape functions is satisfied in fourth-order phase-field model. Two different consistent integration schemes with high efficiency and high accuracy, namely quadratically consistent one-point integration scheme and three-point integration scheme, are employed to evaluate the fourth-order phase-field equation and displacement equation, respectively. To reduce computational cost, the adaptive strategy of local background mesh refinement is established on basis of strain energy history and phase field. Background mesh is refined via the insertion of nodes at the midpoints of each side of the integration element. Comparing with EFG method, computational accuracy and efficiency are enhanced by the proposed method. Load-displacement response, the smoothness of the resulting stress and crack paths are predicted well.

MTH subdivision scheme with sharp and semi-sharp features

In this paper, a new non-stationary bivariate subdivision scheme is proposed to obtain a more flexible limit surface. By rewriting the univariate subdivision scheme (Fakhar et al., 2022) under two different methods, the tensor product and the subdivision rules for extraordinary points/faces are constructed. The new scheme produces continuous limit surfaces of higher order, except at extraordinary points/faces where the continuity is G1. Algorithms for generating sharp and semi-sharp features on the hybrid subdivision surface are presented. Numerical examples are also given to show the performance of these new schemes.

Towards an error indicator-based [formula omitted]-adaptive refinement scheme in kinematic upper-bound limit analysis with the presence of seepage forces

This paper presents a new computational strategy for kinematic upper bound limit analysis in the presence of seepage forces with an improved mesh refinement scheme. In particular, the original adaptive refinement scheme is enhanced with a simple but efficient error-indicator of the nodal plastic dissipation for high-order elements. The proposed refinement criteria facilitate a precise quantitative evaluation of errors in the nodal plastic energy dissipation across each element and ensure that the rate of total refined element number gradually decreases with adaptive step. Adhering to two-dimensional steady state seepage condition, numerical details regarding the calculation of total water head distributions for the seepage field are provided. In a similar manner as treating the unit weight of the soil, the effects of seepage forces are incorporated as body forces in the upper bound formulation. Detailed numerical procedure of the proposed error indicator-based h-adaptive refinement scheme incorporating the inclusion of seepage forces is addressed and implemented in the in-house code. Several benchmark problems, including both homogeneous and non-homogeneous soils, are analyzed to evaluate the excellent performance of the proposed error indicator-based h-adaptive refinement scheme in upper-bound limit analysis and assess the influence of seepage forces on the stability of geostructures through comparison with that for dry condition.

FEM/Wideband FMBEM coupling based on subdivision isogeometry for structural-acoustic design sensitivity analysis

A computer simulation approach known as the isogeometric (IGA) method may directly use the surface information of geometric model. In 3D computer graphics, Loop subdivision surfaces are a common method for creating complicated shapes. In this study, we propose a coupling algorithm that utilizes Loop subdivision surfaces and a direct differentiation method for the computation of acoustic-fluid-structure interaction and the performance of structural-acoustic sensitivity analysis. This algorithm combines the finite element method (FEM) and wideband fast multipole boundary element method (FMBEM). Because of that the proposed method is of a great ability of integrating the numerical calculation and computer-aided modeling, the current technique can deliver results quickly and accurately. The numerical prediction of the effects of vibrating structures with arbitrary shape within sound field is made feasible by the FEM/Wideband FMBEM technique. Calculation examples are provided to show the applicability and effectiveness of the suggested method.

Screening through a broad pool: Towards better diversity for lexically constrained text generation

Lexically constrained text generation (CTG) is to generate text that contains given constrained keywords. However, the text diversity of existing models is still unsatisfactory. In this paper, we propose a lightweight dynamic refinement strategy that aims at increasing the randomness of inference to improve generation richness and diversity while maintaining a high level of fluidity and integrity. Our basic idea is to enlarge the number and length of candidate sentences in each iteration, and choose the best for subsequent refinement. On the one hand, different from previous works, which carefully insert one token between two words per action, we insert an uncertain number of tokens following a well-designed distribution. To ensure high-quality decoding, the insertion number increases as more words are generated. On the other hand, we randomly mask an increasing number of generated words to force Pre-trained Language Models (PLMs) to examine the whole sentence via reconstruction. We have conducted extensive experiments and designed four dimensions for human evaluation. Compared with important baseline (CBART (He, 2021)), our method improves the 1.3% (B-2), 0.1% (B-4), 0.016 (N-2), 0.016 (N-4), 5.7% (M), 1.9% (SB-4), 0.6% (D-2), 0.5% (D-4) on One-Billion-Word dataset (Chelba et al., 2014) and 1.6% (B-2), 0.1% (B-4), 0.121 (N-2), 0.120 (N-4), 0.0% (M), 6.7% (SB-4), 2.7% (D-2), 3.8% (D-4) on Yelp dataset (Cho et al., 2018). The results demonstrate that our method is more diverse and plausible.

DYS-SLAM: A real-time RGBD SLAM combined with optical flow and semantic information in a dynamic environment1

Traditional Simultaneous Localization and Mapping application in dynamic situations is constrained by static assumptions. However, the majority of well-known dynamic SLAM systems use deep learning to identify dynamic objects, which creates the issue of trade-offs between accuracy and real-time. To tackle this issue, this work suggests a unique dynamic semantics method(DYS-SLAM) for semantic simultaneous localization and mapping that strikes a trade-off between high accuracy and high real-time performance. We propose M-LK, an enhanced Lucas-Kanade(LK) optical flow method. This technique keeps the continuous motion and greyscale consistency assumptions from the original method while switching out the spatial consistency assumption for a motion consistency assumption, reducing sensitivity to image gradients to identify dynamic feature points and regions efficiently. In order to increase segmentation accuracy while maintaining real-time performance, we develop a segmentation refinement scheme that projects 3D point cloud segmentation results into 2D object detection zones. A dense semantic octree graph is built in the interim to expedite the high-level process. Compared to the four equivalent dynamic SLAM approaches, experiments on the publicly available TUM RGB-D dataset demonstrate that the DYS-SLAM method offers competitive localization accuracy and satisfactory real-time performance in both high and low-dynamic scenarios.