Multimodal Large Language Models (MLLMs) have demonstrated impressive capabilities in vision-language understanding. Recently, with the integration of test-time scaling techniques, these models have also shown strong potential in visual reasoning. However, most existing reasoning approaches remain text-level in nature: MLLMs are prompted to explore various combinations of textual tokens via their underlying language model, while the visual input remains fixed throughout the reasoning process. This paradigm limits the model’s ability to fully exploit rich visual information, particularly when dealing with images containing numerous fine-grained elements. In such cases, vision-level reasoning becomes crucial—where models dynamically zoom into specific regions of the image to gather detailed visual cues necessary for accurate decision-making. In this paper, we propose Zoom Eye, a training-free, model-agnostic tree search algorithm tailored for vision-level reasoning. Zoom Eye treats an image as a hierarchical tree structure, where each child node represents a zoomed-in sub-region of its parent, and the root corresponds to the full image. The algorithm enables MLLMs to simulate human-like zooming behavior by navigating from root to leaf nodes in search of task-relevant visual evidence. We experiment on a series of elaborate high-resolution benchmarks and the results demonstrate that Zoom Eye not only consistently improves the performance of a series of MLLMs with large margin (e.g., InternVL2.5-8B increases by 15.71% and 17.69% on HR-Bench) but also enables small 3-8B MLLMs to outperform strong large models such as GPT-4o.

Aiming at the problem that the traffic flow prediction model does not consider the road contextual correlation and implicit spatial dependency, a spatio-temporal traffic flow prediction model based on adaptive graph convolution (Transformer-Graph Convolutional Recurrent Network, TGCRN) is proposed, which extracts spatio-temporal features by using an encoder-decoder structure. The encoder part designs a convolutional multi-head self-attention mechanism to extract spatio-temporal features from both local and global perspectives, and uses a traffic pattern memory to augment the node’s pattern information to assist future prediction. The decoder part injects an adaptive graph convolution module into the recurrent neural network to capture both temporal and spatial dependencies and outputs future traffic scenarios. The model extracts the hidden spatial dependencies from the unknown graph structure without any a priori knowledge guidance and is able to model complex and dynamic spatio-temporal relationships. Extensive experiments based on four real traffic datasets confirm the effectiveness of our approach.

The long-tail problem presents a significant challenge to the advancement of semantic segmentation in ultra-high-resolution (UHR) satellite imagery. While previous efforts in UHR semantic segmentation have largely focused on multibranch network architectures that emphasize multi-scale feature extraction and fusion, they have often overlooked the importance of addressing the long-tail issue. In contrast to prior UHR methods that focused on independent feature extraction, we emphasize data augmentation and multimodal feature fusion to alleviate the long-tail problem. In this paper, we introduce SRMF, a novel framework for semantic segmentation in UHR satellite imagery. Our approach addresses the long-tail class distribution by incorporating a multi-scale cropping technique alongside a data augmentation strategy based on semantic reordering and resampling. To further enhance model performance, we propose a multimodal fusion-based general representation knowledge injection method, which, for the first time, fuses text and visual features without the need for individual region text descriptions, extracting more robust features. Extensive experiments on the URUR, GID, and FBP datasets demonstrate that our method improves mIoU by 3.33%,0.66%, and 0.98%, respectively, achieving state-of-the-art performance. Code is available at: https://github.com/BinSpa/SRMF.git.

Pretrained vision-language models (VLMs) utilizing extensive image–text paired data have demonstrated unprecedented image–text association capabilities, achieving remarkable results across various downstream tasks. A critical challenge is how to make use of existing large-scale pretrained VLMs, which are trained on common objects, to perform the domain-specific transfer for accomplishing domain-related downstream tasks. In this article, we present an image–text paired dataset in the field of remote sensing (RS), RS5M, which has 5 million RS images with English descriptions. The dataset is obtained from filtering publicly available image–text paired datasets and captioning label-only RS datasets with pretrained VLM. These constitute the first large-scale RS image–text paired dataset. Additionally, we present GeoRSCLIP by fine-tuning (FT) or applying parameter-efficient FT (PEFT) methods to the CLIP model using RS5M. Experimental results show that our proposed dataset is highly effective for various tasks, and our model GeoRSCLIP improves upon the baseline or previous state-of-the-art model by 3%–20% in zero-shot classification (ZSC) tasks, 3%–6% in RS cross-modal text–image retrieval (RSCTIR) and 4%–5% in semantic localization (SeLo) tasks. Dataset and models have been released in:https://github.com/om-ai-lab/RS5M.

Most graph-network-based meta-learning approaches model instance-level relation of examples. We extend this idea further to explicitly model the distribution-level relation of one example to all other examples in a 1-vs-N manner. We propose a novel approach named distribution propagation graph network (DPGN) for few-shot learning. It conveys both the distribution-level relations and instance-level relations in each few-shot learning task. To combine the distribution-level relations and instance-level relations for all examples, we construct a dual complete graph network which consists of a point graph and a distribution graph with each node standing for an example. Equipped with dual graph architecture, DPGN propagates label information from labeled examples to unlabeled examples within several update generations. In extensive experiments on few-shot learning benchmarks, DPGN outperforms state-of-the-art results by a large margin in 5%~12% under supervised setting and 7%~13% under semi-supervised setting. Code will be released.