Recently, deep learning networks have demonstrated strong performance in hyperspectral image (HSI) classification tasks. 2D Convolutional Neural Networks (2D-CNNs) perform well in the spatial dimension, but tend to overlook spectral information. Meanwhile, 3D Convolutional Neural Networks (3D-CNNs) perform well in the spectral dimension, but are inadequate for the exploitation of spatial and spectral information. To address this issue, a cross-fusion network (CFN) is proposed for HSI classification. Rather than simply combining 2D and 3D convolutions, CFN introduces two dedicated modules: a cross-feature module (CFM) and an adaptive cross-fusion module (ACFM). The CFM performs bidirectional cross-dimensional feature exchange between the 2D and 3D branches to capture local spatial-spectral fusion representations. The ACFM further models long-range pixel dependencies by conducting cross-branch contextual interaction and cross-residual enhancement. Based on these two modules, CFN can jointly exploit local spatial-spectral representations and long-range contextual dependencies. Experiments on the Indian Pines, Houston2013, and Pavia University datasets demonstrate that CFN achieves competitive performance compared with state-of-the-art methods.
In a previous work, the theoretical existence of uncertainty relations governing chromatic opponency for general chromatic states within a quantum information-based model of color perception has been established. In the present study, we extend this result by deriving uncertainty relations for chromatic states perceived by a human observer, modeled as a quantum effect. Our analysis indicates that, once the presence of an observer is explicitly taken into account, only uncertainty relations associated with highly saturated stimuli are likely to be experimentally detectable.
Dense video captioning is a multimodal task that involves visual feature extraction and natural language generation. Previous approaches face several challenges, such as reliance on complex "localize-then-describe" frameworks, limited contextual information, and the potential for inaccurate detection. In this paper, we propose an end-to-end dense video captioning model that leverages multimodal data and multi-level visual features, based on the pretrained ViFi-CLIP model. Our approach constructs an integrated network that unifies video segmentation and caption generation into a single end-to-end framework, effectively utilizing visual, audio, and textual modalities. In selecting visual features, we comprehensively consider the impact of grid and region-based features and employ ViFi-CLIP to map these features into a joint vision–language embedding space, thereby enhancing the semantic relevance of the visual representations. Extensive experiments conducted on the ActivityNet dataset demonstrate the effectiveness of our model. Results show that our approach outperforms PDVC by 8–10
Camera–LiDAR fusion has shown great promise in advancing road detection for autonomous driving, combining the semantic richness of RGB imagery with the depth accuracy of LiDAR. However, its practical deployment faces two main challenges: (1) the high cost and limited accessibility of LiDAR sensors hinder their large-scale adoption, and (2) the scarcity of labeled data in unseen target domains limits the generalization capability of supervised methods. To overcome these limitations, we propose SPADA-Road, an unsupervised framework that integrates Superpixel-guided segmentation with a novel Pseudo-LiDAR (PL) generation module. Our PL module synthesizes depth cues from monocular RGB inputs, effectively augmenting training data without requiring physical LiDAR sensors. To enable robust cross-domain generalization, we adopt an adversarial domain adaptation strategy that aligns feature distributions between labeled source and unlabeled target domains. We evaluate SPADA-Road on the KITTI Road dataset for training, and validate its performance on two LiDAR-free benchmarks—Cityscapes and CamVid. Extensive experiments demonstrate that our method achieves superior performance compared to several state-of-the-art baselines, highlighting its effectiveness in LiDAR-free and label-scarce scenarios.
We present a comparative study of segmentation methods for high-power laser applications, focusing on two specific challenges: detection of microscopic surface damage on optical components and detection of radiochromic films for reconstructing high-dimensional particle phase space distributions. Both applications involve complex morphological variations and non-homogeneous contrast conditions, requiring robust and scalable analysis methods. We evaluate two conventional algorithms and two deep learning-based instance segmentation models, including YOLOv8n-seg and a Detectron2-based Mask R-CNN implementation. All models are evaluated on real datasets that reflect the experimental complexities. We focus particular attention to the accuracy of contour detection, using geometric evaluation metrics such as radial contour comparison, Hausdorff distance, Chamfer distance, as well as intersection-over-union, and analysing runtime performance. Our results indicate that the YOLOv8n-seg model outperforms the conventional surface damage segmentation method in accuracy, but with 12 times higher computational requirements. In contrast, for radiochromic films analysing YOLOv8n-seg achieves both higher accuracy and faster evaluation. In comparison to YOLOv8n-seg model, Detectron2-based Mask R-CNN implementation lags in both segmentation performance and runtime. These results highlight the potential of YOLOv8n-seg model in addressing specific data-related challenges in modern laser diagnostics and support their role in the development of next-generation automated analysis systems.
