CP-DETR: Concept Prompt Guide DETR Toward Stronger Universal Object Detection

Recent research on universal object detection aims to introduce language in a SoTA closed-set detector and then generalize the open-set concepts by constructing large-scale (text-region) datasets for training. However, these methods face two main challenges: (i) how to efficiently use the prior information in the prompts to genericise objects and (ii) how to reduce alignment bias in the downstream tasks, both leading to sub-optimal performance in some scenarios beyond pre-training. To address these challenges, we propose a strong universal detection foundation model called CP-DETR, which is competitive in almost all scenarios, with only one pre-training weight. Specifically, we design an efficient prompt visual hybrid encoder that enhances the information interaction between prompt and visual through scale-by-scale and multi-scale fusion modules. Then, the hybrid encoder is facilitated to fully utilize the prompted information by prompt multi-label loss and auxiliary detection head. In addition to text prompts, we have designed two practical concept prompt generation methods, visual prompt and optimized prompt, to extract abstract concepts through concrete visual examples and stably reduce alignment bias in downstream tasks. With these effective designs, CP-DETR demonstrates superior universal detection performance in a broad spectrum of scenarios. For example, our Swin-T backbone model achieves 47.6 zero-shot AP on LVIS, and the Swin-L backbone model achieves 32.2 zero-shot AP on ODinW35. Furthermore, our visual prompt generation method achieves 68.4 AP on COCO val by interactive detection, and the optimized prompt achieves 73.1 fully-shot AP on ODinW13.

Multi-Modal Environmental Sensing Based Path Loss Prediction for V2I Communications

The stability and reliability of wireless data transmission in vehicular networks face significant challenges due to the high dynamics of path loss caused by the complexity of rapidly changing environments. This paper proposes a multi-modal environmental sensing-based path loss prediction architecture (MES-PLA) for V2I communications. First, we establish a multi-modal environment data and channel joint acquisition platform to generate a spatio-temporally synchronized and aligned dataset of environmental and channel data. Then we designed a multi-modal feature extraction and fusion network (MFEF-Net) for multi-modal environmental sensing data. MFEF-Net extracts features from RGB images, point cloud data, and GPS information, and integrates them with an attention mechanism to effectively leverage the strengths of each modality. The simulation results demonstrate that the Root Mean Square Error (RMSE) of MES-PLA is 2.20 dB, indicating a notable improvement in prediction accuracy compared to single-modal sensing data input. Moreover, MES-PLA exhibits enhanced stability under varying illumination conditions compared to single-modal methods.

Holmes-VAU: Towards Long-term Video Anomaly Understanding at Any Granularity

How can we enable models to comprehend video anomalies occurring over varying temporal scales and contexts? Traditional Video Anomaly Understanding (VAU) methods focus on frame-level anomaly prediction, often missing the interpretability of complex and diverse real-world anomalies. Recent multimodal approaches leverage visual and textual data but lack hierarchical annotations that capture both short-term and long-term anomalies. To address this challenge, we introduce HIVAU-70k, a large-scale benchmark for hierarchical video anomaly understanding across any granularity. We develop a semi-automated annotation engine that efficiently scales high-quality annotations by combining manual video segmentation with recursive free-text annotation using large language models (LLMs). This results in over 70,000 multi-granular annotations organized at clip-level, event-level, and video-level segments. For efficient anomaly detection in long videos, we propose the Anomaly-focused Temporal Sampler (ATS). ATS integrates an anomaly scorer with a density-aware sampler to adaptively select frames based on anomaly scores, ensuring that the multimodal LLM concentrates on anomaly-rich regions, which significantly enhances both efficiency and accuracy. Extensive experiments demonstrate that our hierarchical instruction data markedly improves anomaly comprehension. The integrated ATS and visual-language model outperform traditional methods in processing long videos. Our benchmark and model are publicly available at https://github.com/pipixin321/HolmesVAU.

Relevance-guided Audio Visual Fusion for Video Saliency Prediction

Audio data, often synchronized with video frames, plays a crucial role in guiding the audience's visual attention. Incorporating audio information into video saliency prediction tasks can enhance the prediction of human visual behavior. However, existing audio-visual saliency prediction methods often directly fuse audio and visual features, which ignore the possibility of inconsistency between the two modalities, such as when the audio serves as background music. To address this issue, we propose a novel relevance-guided audio-visual saliency prediction network dubbed AVRSP. Specifically, the Relevance-guided Audio-Visual feature Fusion module (RAVF) dynamically adjusts the retention of audio features based on the semantic relevance between audio and visual elements, thereby refining the integration process with visual features. Furthermore, the Multi-scale feature Synergy (MS) module integrates visual features from different encoding stages, enhancing the network's ability to represent objects at various scales. The Multi-scale Regulator Gate (MRG) could transfer crucial fusion information to visual features, thus optimizing the utilization of multi-scale visual features. Extensive experiments on six audio-visual eye movement datasets have demonstrated that our AVRSP network achieves competitive performance in audio-visual saliency prediction.

Multi-task Feature Enhancement Network for No-Reference Image Quality Assessment

Due to the scarcity of labeled samples in Image Quality Assessment (IQA) datasets, numerous recent studies have proposed multi-task based strategies, which explore feature information from other tasks or domains to boost the IQA task. Nevertheless, multi-task strategies based No-Reference Image Quality Assessment (NR-IQA) methods encounter several challenges. First, existing methods have not explicitly exploited texture details, which significantly influence the image quality. Second, multi-task methods conventionally integrate features through simple operations such as addition or concatenation, thereby diminishing the network's capacity to accurately represent distorted features. To tackle these challenges, we introduce a novel multi-task NR-IQA framework. Our framework consists of three key components: a high-frequency extraction network, a quality estimation network, and a distortion-aware network. The high-frequency extraction network is designed to guide the model's focus towards high-frequency information, which is highly related to the texture details. Meanwhile, the distortion-aware network extracts distortion-related features to distinguish different distortion types. To effectively integrate features from different tasks, a feature fusion module is developed based on an attention mechanism. Empirical results from five standard IQA databases confirm that our method not only achieves high performance but also exhibits robust generalization ability.

High-Frequency Enhanced Hybrid Neural Representation for Video Compression

Neural Representations for Videos (NeRV) have simplified the video codec process and achieved swift decoding speeds by encoding video content into a neural network, presenting a promising solution for video compression. However, existing work overlooks the crucial issue that videos reconstructed by these methods lack high-frequency details. To address this problem, this paper introduces a High-Frequency Enhanced Hybrid Neural Representation Network. Our method focuses on leveraging high-frequency information to improve the synthesis of fine details by the network. Specifically, we design a wavelet high-frequency encoder that incorporates Wavelet Frequency Decomposer (WFD) blocks to generate high-frequency feature embeddings. Next, we design the High-Frequency Feature Modulation (HFM) block, which leverages the extracted high-frequency embeddings to enhance the fitting process of the decoder. Finally, with the refined Harmonic decoder block and a Dynamic Weighted Frequency Loss, we further reduce the potential loss of high-frequency information. Experiments on the Bunny and UVG datasets demonstrate that our method outperforms other methods, showing notable improvements in detail preservation and compression performance.

ChannelGPT: A Large Model to Generate Digital Twin Channel for 6G Environment Intelligence

6G is envisaged to provide multimodal sensing, pervasive intelligence, global coverage, global coverage, etc., which poses extreme intricacy and new challenges to the network design and optimization. As the core part of 6G, wireless channel is the carrier and enabler for the flourishing technologies and novel services, which intrinsically determines the ultimate system performance. However, how to describe and utilize the complicated and high-dynamic characteristics of wireless channel accurately and effectively still remains great hallenges. To tackle this, digital twin is envisioned as a powerful technology to migrate the physical entities to virtual and computational world. In this article, we propose a large model driven digital twin channel generator (ChannelGPT) embedded with environment intelligence (EI) to enable pervasive intelligence paradigm for 6G network. EI is an iterative and interactive procedure to boost the system performance with online environment adaptivity. Firstly, ChannelGPT is capable of utilization the multimodal data from wireless channel and corresponding physical environment with the equipped sensing ability. Then, based on the fine-tuned large model, ChannelGPT can generate multi-scenario channel parameters, associated map information and wireless knowledge simultaneously, in terms of each task requirement. Furthermore, with the support of online multidimensional channel and environment information, the network entity will make accurate and immediate decisions for each 6G system layer. In practice, we also establish a ChannelGPT prototype to generate high-fidelity channel data for varied scenarios to validate the accuracy and generalization ability based on environment intelligence.

Wireless Environment Information Sensing, Feature, Semantic, and Knowledge: Four Steps Towards 6G AI-Enabled Air Interface

The air interface technology plays a crucial role in optimizing the communication quality for users. To address the challenges brought by the radio channel variations to air interface design, this article proposes a framework of wireless environment information-aided 6G AI-enabled air interface (WEI-6G AI$^{2}$), which actively acquires real-time environment details to facilitate channel fading prediction and communication technology optimization. Specifically, we first outline the role of WEI in supporting the 6G AI$^{2}$ in scenario adaptability, real-time inference, and proactive action. Then, WEI is delineated into four progressive steps: raw sensing data, features obtained by data dimensionality reduction, semantics tailored to tasks, and knowledge that quantifies the environmental impact on the channel. To validate the availability and compare the effect of different types of WEI, a path loss prediction use case is designed. The results demonstrate that leveraging environment knowledge requires only 2.2 ms of model inference time, which can effectively support real-time design for future 6G AI$^{2}$. Additionally, WEI can reduce the pilot overhead by 25\%. Finally, several open issues are pointed out, including multi-modal sensing data synchronization and information extraction method construction.

Bridging Domain Gap of Point Cloud Representations via Self-Supervised Geometric Augmentation

Recent progress of semantic point clouds analysis is largely driven by synthetic data (e.g., the ModelNet and the ShapeNet), which are typically complete, well-aligned and noisy free. Therefore, representations of those ideal synthetic point clouds have limited variations in the geometric perspective and can gain good performance on a number of 3D vision tasks such as point cloud classification. In the context of unsupervised domain adaptation (UDA), representation learning designed for synthetic point clouds can hardly capture domain invariant geometric patterns from incomplete and noisy point clouds. To address such a problem, we introduce a novel scheme for induced geometric invariance of point cloud representations across domains, via regularizing representation learning with two self-supervised geometric augmentation tasks. On one hand, a novel pretext task of predicting translation distances of augmented samples is proposed to alleviate centroid shift of point clouds due to occlusion and noises. On the other hand, we pioneer an integration of the relational self-supervised learning on geometrically-augmented point clouds in a cascade manner, utilizing the intrinsic relationship of augmented variants and other samples as extra constraints of cross-domain geometric features. Experiments on the PointDA-10 dataset demonstrate the effectiveness of the proposed method, achieving the state-of-the-art performance.

MSVM-UNet: Multi-Scale Vision Mamba UNet for Medical Image Segmentation

State Space Models (SSMs), especially Mamba, have shown great promise in medical image segmentation due to their ability to model long-range dependencies with linear computational complexity. However, accurate medical image segmentation requires the effective learning of both multi-scale detailed feature representations and global contextual dependencies. Although existing works have attempted to address this issue by integrating CNNs and SSMs to leverage their respective strengths, they have not designed specialized modules to effectively capture multi-scale feature representations, nor have they adequately addressed the directional sensitivity problem when applying Mamba to 2D image data. To overcome these limitations, we propose a Multi-Scale Vision Mamba UNet model for medical image segmentation, termed MSVM-UNet. Specifically, by introducing multi-scale convolutions in the VSS blocks, we can more effectively capture and aggregate multi-scale feature representations from the hierarchical features of the VMamba encoder and better handle 2D visual data. Additionally, the large kernel patch expanding (LKPE) layers achieve more efficient upsampling of feature maps by simultaneously integrating spatial and channel information. Extensive experiments on the Synapse and ACDC datasets demonstrate that our approach is more effective than some state-of-the-art methods in capturing and aggregating multi-scale feature representations and modeling long-range dependencies between pixels.