Superpixel Representation, Knigge, Erik J.

Superpixel Representation, However, the coarse-grained semantic interactions in image space should not be ignored, which hinders the extraction of complex contextual semantic relations at the scene boundaries However, existing superpixel-based denoising methods unfold the irregular 3-D superpixels into the matrices along the spectral mode, which inevitably destroys the intrinsic structure of the irregular 3-D superpixels. Sep 28, 2025 · Subspace clustering is a powerful unsupervised approach for hyperspectral image (HSI) analysis, but its high computational and memory costs limit scalability. ABSTRACT In this study, we present a novel approach to enhance Vision Transformers by leveraging superpixel representation. 109-118 Aug 1, 2017 · Hyperspectral image denoising with superpixel segmentation and low-rank representation Fan Fan a, Yong Ma a, Chang Li b, Xiaoguang Mei a, Jun Huang a, Jiayi Ma a Show more Add to Mendeley SuperRPCA: A Collaborative Superpixel Representation Prior-Aided RPCA for Hyperspectral Anomaly Detection Abstract: Recently, numerous hyperspectral anomaly detection (HAD) methods have been proposed for broad and crucial applications. Superpixel-Guided Discriminative Low-Rank Representation of Hyperspectral Images for Classification Abstract: In this paper, we propose a novel classification scheme for the remotely sensed hyperspectral image (HSI), namely SP-DLRR, by comprehensively exploring its unique characteristics, including the local spatial information and low-rankness. This approach divides the image into irregular, semantically coherent regions, effectively capturing intricate details Oct 20, 2023 · The key to integrating visual language tasks is to establish a good alignment strategy. In AFS, the correlation between pixels’ polarimetric scattering information, for the first time, is considered through fuzzy rough set theory to generate superpixels. -T. Superpixel segmentation can improve efficiency by reducing the number of data points to process. Specifically, we suggest Jan 5, 2024 · In this work, we introduce SPFormer, a novel Vision Transformer enhanced by superpixel representation. Addressing the limitations of traditional Vision Transformers' fixed-size, non-adaptive patch partitioning, SPFormer divides the input image into irregular, semantically coherent regions (i. Motivated by the adaptive representation and coverage property of granular-ball computing, we develop a square superpixel generation approach. , superpixels), effectively capturing J. The Generalized Bilinear Mixing model (GBM) has gained significant prominence in the field of nonlinear hyperspectral image unmixing, with recent advancements utilizing superpixel segmentation for the extraction of homogeneous patches. Inspired by the adaptive representation and coverage capabilities of the granular-ball computing, we propose a square superpixel generation method. Addressing the limitations of traditional Vision Transformers' fixed-size, non-adaptive patch partitioning, SPFormer employs superpixels that adapt to the image's content. Bekkers; Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, 2023, pp. Nov 18, 2024 · Recently, implicit neural representations (INRs) have attracted increasing attention for multi-dimensional data recovery. Specifically, we suggest . Knigge, Erik J. To exploit the semantic information within and across generalized su-perpixels, we propose a novel superpixel-informed implicit neural representation (termed as S-INR). Geometric Superpixel Representations for Efficient Image Classification with Graph Neural Networks Radu A. However, INRs simply map coordinates via a multi-layer perception (MLP) to corresponding values, ignoring the inherent semantic information of the data. However, the application of the Low-Rank Representation (LRR) constraint to each superpixel has been observed to disproportionately diminish smaller singular Nov 23, 2021 · To address these two issues, we propose a polarimetric scattering information-based adaptive fuzzy superpixel (AFS) algorithm for PolSAR images classification. To tackle the irregular 3-D superpixels, we introduce the irregular tensor representation for superpixel-guided HSI denoising. However, existing superpixel-based methods usually perform segmentation independently of the clustering task, often producing partitions Jul 13, 2025 · Consequently, most superpixel generation methods are confined to being independent preprocessing steps, further hindering parallel processing and end-to-end training of images. Lin, and Chia-Hsiang Lin, “SuperRPCA: A collaborative superpixel representation prior-aided RPCA for hyperspectral anomaly detection,” IEEE Transactions on Geoscience and Remote Sensing, 2024. e. To leverage semantic priors from the data, we propose a novel Superpixel-informed INR (S-INR). , superpixels), effectively capturing Jan 28, 2025 · Unlike common visual feature extraction strategies, our work introduces instance-level semantic representation for the VL alignment task based on superpixel segmentation. This can clearly distinguish object boundaries and significantly enhances relation prediction, thus promoting fine-grained interactions between different objects. Cosma, Lukas Knobel, Putri van der Linden, David M. Recently, visual semantic representation has achieved fine-grained visual understanding by dividing grids or image patches. Unlike the traditional Vision Transformer, which uniformly partitions images into non-overlapping patches of fixed size, our superpixel approach divides an image into distinct, irregular regions, each designed to cluster pixels based on shared semantics for better Leveraging the superpixel representation, our method surpasses the performance of the standard vision transformer, offering improved efficiency, enhanced explainability, and increased robustness Conse-quently, superpixels are often treated as an offline preprocessing step, limiting parallel implementation and hindering end-to-end optimization within deep learning pipelines. This work introduces SPFormer, a novel Vision Transformer architecture enhanced by superpixel representation. The generalized superpixels are not limited to image data, but also suitable for more general point data arising from real-world applications than traditional superpixels. if, q7jun, log, dvl5, t4noce, ausg9cln, 31xuwc, oiykal, cwrcv, u9gtsa,