Change captioning is a task that describes changes in image pairs using natural language. This task is more complex than single-image captioning as it requires a comprehensive understanding of each image and the ability to recognize and describe the semantic changes in image pairs. The key challenge lies in making the network generate an accurate and stable change representation under the interference of viewpoint shift. In this paper, we propose a cross-view and multi-step interaction network to generate robust change representation to resist pseudo-change. Specifically, in the intra-image representation learning stage, a cross-view interaction encoder is designed to enhance internal relationships by cross-referencing in image pairs. In the change feature learning stage, a multi-step change perceptron is employed to capture the change semantics from coarse to fine progressively. Then, a fusion module dynamically combines them as a fine-grained change representation. Besides, we propose a backward representation reconstruction module that facilitates the capture of semantic changes, thus improving the quality of captions in a self-supervised manner. Extensive experiments have shown that the method effectively captures real semantic changes under the interference of viewpoint shift and achieves state-of-the-art performance on five public datasets. The code is available at https://github.com/TTXiann/CVMSI

Recent memory-based approaches have shown promising performance in semi-supervised video object segmentation. However, they still face two core challenges. First, their heavy reliance on pixel-level feature matching makes them vulnerable to background distractors that closely resemble the target object, resulting in erroneous segmentations. Second, these incorrect segmentation masks are indiscriminately stored in memory, causing error accumulation and deteriorating performance over time. To address these challenges, we propose the Instance Aware Self-Correcting Model (IASC), which integrates three key components: the Dual-Stream Framework, the Online Adaptation Module (OAM), and the Self-Correcting Module (SCM). The Dual-Stream Framework leverages complementary representations by combining the Appearance Feature Stream, which recovers fine-grained pixel-level details through similarity-based matching, and the Semantic Mask Stream, which generates instance-level attention using reference masks to suppress distractors and enhance segmentation robustness. The OAM dynamically fine-tunes the key projector using the first annotated frame, improving the accuracy of pixel-level matching and the reliability of instance-level guidance. Additionally, the SCM refines unreliable predictions, effectively mitigating error propagation during memory updates. Extensive experiments on benchmark datasets demonstrate that IASC achieves state-of-the-art performance, striking an effective balance between segmentation accuracy and computational efficiency.

In recent years, Transformer-based models have achieved remarkable success in time series forecasting. To address the limitations of point-wise information and reduce computational complexity, several patch-wise Transformer models have been proposed. However, most sparse attention mechanisms between patches are static, often leading to information loss. To overcome this issue, we propose BiTST, a novel time series forecasting model that leverages bi-level routing attention along a single dimension (1D-BRA). The model is built on a Transformer encoder framework and employs a multilayer pyramid structure to reduce computational complexity. We also introduce a convolutional embedding layer, enabling flexible patch segmentation. The instance-level 1D-BRA effectively filters out irrelevant key-value pairs at a coarse level, achieving sparse attention through dynamic query-based matching. Furthermore, we incorporate adaptive Huber loss into the training process to robustly handle outliers during early training stages, while ensuring high accuracy in later stages. This combination enhances the model’s overall robustness. Extensive experiments on diverse benchmarks demonstrate that BiTST achieves superior accuracy and high efficiency across a wide range of tasks in time series forecasting.

Remote Sensing Image Change Captioning (RSICC) plays a critical role in automated environmental monitoring by generating natural language descriptions that provide intuitive interpretations of changes between bi-temporal remote sensing images. Despite recent advancements, existing methods suffer from two fundamental limitations: (1) static encoding architectures fail to account for scene complexity and semantic diversity during feature extraction, leading to inflexible representation learning; and (2) symmetric computational structures are inherently unsuitable for modeling the asymmetric temporal dependencies inherent in “before-to-after” image pairs. To address these challenges, we propose a Dynamic Asymmetric Encoder (DAE), which introduces two key innovations. First, we design a difference-guided dynamic convolution module that adaptively adjusts convolutional parameters using input-driven scaling factors and offsets, thereby enabling scene-aware intra-image feature enhancement. Second, we develop a Multi-expert Temporal Interaction (METI) module that establishes an asymmetric computational topology: the “after” image branch actively perceives change information relative to the “before” image through three heterogeneous interaction experts, followed by feature fusion. This design allocates greater computational capacity to the “after” image branch while preserving temporal coherence. Furthermore, we introduce a Multi-level Feature Aggregator (MFA) that enhances the representation of salient changed regions across multiple scales via an iterative reinforcement mechanism. Experimental results validate the effectiveness of the proposed method, demonstrating state-of-the-art performance on two benchmark RSICC datasets. We publicly release our code repository at https://github.com/TTXiann/Dynamic-Asymmetric to facilitate future research.

Unsupervised Video Object Segmentation (UVOS) aims to autonomously recognize and segment primary foreground objects within a given video sequence without additional inputs. Current methods often rely on motion information, such as optical flow, between adjacent frames, which is fused with appearance features. However, this approach overlooks the latent deep object representations within the features and neglects the inter-frame connectivity over longer distances. In this paper, we present a pioneering solution, the Multi-Involution Memory Network (MIMN), designed to overcome these challenges. Specifically, we introduce the Multi-Involution Selective Kernel (MISK), a mechanism for aggregating deep representation features across various receptive fields. To facilitate contextual information exchange between frames at different distances, our proposed Inter Memory Aggregation (IMA) incorporates a novel temporal attention mechanism. This mechanism selectively identifies the most crucial features in the memory bank, enhancing the prediction accuracy for the current frame. Extensive empirical studies conducted on multiple benchmarks demonstrate the promising performance and high efficiency of the proposed MIMN method. The integration of MISK and IMA contributes to improved UVOS outcomes, addressing limitations present in existing approaches.

Person re-identification (Re-ID) is a crucial technology for intelligent security in Internet of Things (IoT) systems. Recently, unsupervised learning has been widely used for person Re-ID due to its generalization property. However, the effectiveness of commonly used unsupervised clustering methods heavily relies on the quality of the clustered pseudo-labels. Moreover, pedestrian shots in real scenes are prone to factors such as occlusion. In this paper, we propose a novel Global and Local Joint Contrastive Learning (GLCL) framework based on the memory bank. Specifically, we establish separate memory banks for global and local features, which are updated using global simple samples and local hard samples. The GLCL module helps excavate information from simple and hard samples, aiming to overcome the effects of poor retrieval scenarios such as background clutter and occlusion. Additionally, we design an Attributes-Assisted Clustering (AAC)) module that utilizes pedestrian attributes to refine the clustering results. The AAC module can effectively reduce the impact of pseudo-label noise owing to the supplementary information offered by attributes. Our approach shows improved performance in person Re-ID tasks in complex scenarios, providing a promising solution for intelligent security systems in the IoT. Experimental results demonstrate the superiority of our proposed method.

In recent years, many salient object detection (SOD) methods introduce depth cues to boost detection performance in challenging scenes, named as RGB-D SOD. However, how to effectively fuse cross-modal features with various properties (i.e., RGB and depth) has become a key issue that is hard to be avoided. Most existing methods employ simple operations, such as concatenation or summation, for cross-modal fusion, ignoring the negative effects of low-quality depth maps, thus yielding poor performance. In this paper, we design a simple yet effective fusion method, which utilizes 3D convolution to extract modality-specific and modality-shared information respectively for sufficient cross-modal fusion, and combines modality weights to mitigate the interference of invalid information. In addition, we propose a novel multi-level feature integration strategy in the decoder, which explicitly incorporates the low-level detail information and high-level semantic information into the mid-level to generate accurate saliency maps. Extensive experiments on six public datasets show that our method achieves competitive results compared to 17 state-of-the-art methods.

Pedestrians in videos are usually in a moving state, resulting in serious spatial misalignment like scale variations and pose changes, which makes the video-based person re-identification problem more challenging. To address the above issue, in this paper, we propose a Frame-Guided Region-Aligned model (FGRA) for discriminative representation learning in two steps in an end-to-end manner. Firstly, based on a frame-guided feature learning strategy and a non-parametric alignment module, a novel alignment mechanism is proposed to extract well-aligned region features. Secondly, in order to form a sequence representation, an effective feature aggregation strategy that utilizes temporal alignment score and spatial attention is adopted to fuse region features in the temporal and spatial dimensions, respectively. Experiments are conducted on benchmark datasets to demonstrate the effectiveness of the proposed method to solve the misalignment problem and the superiority of the proposed method to the existing video-based person re-identification methods.

Fully Convolutional Neural Networks (FCNs) greatly promote the development of saliency detection. However, most of the FCN-based models have suffered from the structure of salient objects challenges. The extracted multi-scale features by previous models could help locate the objects with various scales, but they cannot contribute to effectively locating the objects with complex shapes, especially the salient regions that might intertwine with non-salient regions. Moreover, the style of decoder in previous models cannot adequately filter out the disturbance in low-level features, which is sub-optimal to sharpen the boundary of salient objects. In this paper, we propose DAGNet that explores the structure of salient objects from multi-level features to precisely detect salient objects. Firstly, the new dense multi-scale context extraction modules (DMCEMs) are implemented to transmit the rich structural information flow of salient objects from shallower layers to deeper layers, by which our model can locate the objects with complex shapes. Secondly, attention-based deeply refining modules (ADRMs) are designed in an effective attention-based style to effectively restore the boundary of objects stage-by-stage. In the style, the semantic information of high-level features is utilized to guide the shallow layer to filter out disturbance and refine the high-level features. Considering the salient objects surrounded by a cluttered scene, we propose a global context extraction module (GCEM) that can sufficiently understand the cluttered scene of an image from a global view. Comprehensive experiments indicate that our model is superior to 13 state-of-the-art models on 5 benchmark datasets under different evaluation metrics.

Nonnegative matrix factorization (NMF) with orthogonality constraints is quite important due to its close relation with the K-means clustering. While existing algorithms for orthogonal NMF impose strict orthogonality constraints, in this letter we propose a penalty method with the aim of performing approximately orthogonal NMF, together with two efficient algorithms respectively based on the Hierarchical Alternating Least Squares (HALS) and the Accelerated Proximate Gradient (APG) approaches. Experimental evidence was provided to show their high efficiency and flexibility by using synthetic and real-world data.