Honggang Chen

Xuyang Liu, Ting Liu, Siteng Huang et al.

Referring expression comprehension (REC) is a vision-language task to locate a target object in an image based on a language expression. Fully fine-tuning general-purpose pre-trained vision-language foundation models for REC yields impressive performance but becomes increasingly costly. Parameter-efficient transfer learning (PETL) methods have shown strong performance with fewer tunable parameters. However, directly applying PETL to REC faces two challenges: (1) insufficient multi-modal interaction between pre-trained vision-language foundation models, and (2) high GPU memory usage due to gradients passing through the heavy vision-language foundation models. To this end, we present M2IST: Multi-Modal Interactive Side-Tuning with M3ISAs: Mixture of Multi-Modal Interactive Side-Adapters. During fine-tuning, we fix the pre-trained uni-modal encoders and update M3ISAs to enable efficient vision-language alignment for REC. Empirical results reveal that M2IST achieves better performance-efficiency trade-off than full fine-tuning and other PETL methods, requiring only 2.11\% tunable parameters, 39.61\% GPU memory, and 63.46\% training time while maintaining competitive performance. Our code is released at https://github.com/xuyang-liu16/M2IST.

Xiaochen Huang, Honggang Chen, Weicheng Zhang et al.

In multimedia application scenarios, images captured under low-illumination conditions often lead to lower accuracy in visual perception tasks compared to those taken in well-lit environments. To tackle this challenge, we propose AMIEOD, an image enhancement-enabled object detection framework for low-illumination scenes, where the two tasks are jointly optimized in a detection performance-oriented manner. Specifically, to fully exploit the information in poorly lit images, a Multi-Experts Image Enhancement Module (MEIEM) is proposed, which leverages diverse enhancement strategies. On this basis, aiming to better align the MEIEM with the detection task, we propose a Detection-Guided Regression Loss (DGRL) that utilizes the detection result to decide the regression target. Moreover, to dynamically select the most suitable enhancement strategy from MEIEM during inference, we construct an Expert Selection Module (ESM) guided by the proposed Detection-Guided Cross-Entropy (DGCE) loss, which formulates the optimization of ESM as a classification task. The improved method is well-matched with current detection algorithms to improve their performance in dim scenes. Extensive experiments on multiple datasets demonstrate that the proposed method significantly improves object detection accuracy in low-illumination conditions. Our code has been released at https://github.com/scujayfantasy/AMIEOD

Yuhang Han, Xuyang Liu, Zihan Zhang et al.

The quadratic complexity of Multimodal Large Language Models (MLLMs) with respect to context length poses significant computational and memory challenges, hindering their real-world deployment. In the paper, we devise a ''filter-correlate-compress'' framework to accelerate the MLLM by systematically optimizing multimodal context length during prefilling. The framework first implements FiCoCo-V, a training-free method operating within the vision encoder. It employs a redundancy-based token discard mechanism that uses a novel integrated metric to accurately filter out redundant visual tokens. To mitigate information loss, the framework introduces a correlation-based information recycling mechanism that allows preserved tokens to selectively recycle information from correlated discarded tokens with a self-preserving compression, thereby preventing the dilution of their own core content. The framework's FiCoCo-L variant further leverages task-aware textual priors to perform token reduction directly within the LLM decoder. Extensive experiments demonstrate that the FiCoCo series effectively accelerates a range of MLLMs, achieves up to 14.7x FLOPs reduction with 93.6% performance retention. Our methods consistently outperform state-of-the-art training-free approaches, showcasing effectiveness and generalizability across model architectures, sizes, and tasks without requiring retraining. Code: https://github.com/kawhiiiileo/FiCoCo

Ting Liu, Xuyang Liu, Siteng Huang et al.

Visual grounding (VG) is a challenging task to localize an object in an image based on a textual description. Recent surge in the scale of VG models has substantially improved performance, but also introduced a significant burden on computational costs during fine-tuning. In this paper, we explore applying parameter-efficient transfer learning (PETL) to efficiently transfer the pre-trained vision-language knowledge to VG. Specifically, we propose \textbf{DARA}, a novel PETL method comprising \underline{\textbf{D}}omain-aware \underline{\textbf{A}}dapters (DA Adapters) and \underline{\textbf{R}}elation-aware \underline{\textbf{A}}dapters (RA Adapters) for VG. DA Adapters first transfer intra-modality representations to be more fine-grained for the VG domain. Then RA Adapters share weights to bridge the relation between two modalities, improving spatial reasoning. Empirical results on widely-used benchmarks demonstrate that DARA achieves the best accuracy while saving numerous updated parameters compared to the full fine-tuning and other PETL methods. Notably, with only \textbf{2.13\%} tunable backbone parameters, DARA improves average accuracy by \textbf{0.81\%} across the three benchmarks compared to the baseline model. Our code is available at \url{https://github.com/liuting20/DARA}.

Xuyang Liu, Ziming Wang, Junjie Chen et al.

Large vision-language models (LVLMs) excel at visual understanding, but face efficiency challenges due to quadratic complexity in processing long multi-modal contexts. While token compression can reduce computational costs, existing approaches are designed for single-view LVLMs and fail to consider the unique multi-view characteristics of high-resolution LVLMs with dynamic cropping. Existing methods treat all tokens uniformly, but our analysis reveals that global thumbnails can naturally guide the compression of local crops by providing holistic context for informativeness evaluation. In this paper, we first analyze dynamic cropping strategy, revealing both the complementary nature between thumbnails and crops, and the distinctive characteristics across different crops. Based on our observations, we propose "Global Compression Commander" (GlobalCom$^2$), a novel plug-and-play token compression framework for HR-LVLMs. GlobalCom$^2$ leverages thumbnail as the "commander" to guide the compression of local crops, adaptively preserving informative details while eliminating redundancy. Extensive experiments show that GlobalCom$^2$ maintains over 90% performance while compressing 90% visual tokens, reducing FLOPs and peak memory to 9.1% and 60%. Our code is available at https://github.com/xuyang-liu16/GlobalCom2.

Daiqiang Li, Zihao Pan, Zeyu Zhang et al.

In recent years, GUI visual agents built upon Multimodal Large Language Models (MLLMs) have demonstrated strong potential in navigation tasks. However, high-resolution GUI screenshots produce a large number of visual tokens, making the direct preservation of complete historical information computationally expensive. In this paper, we conduct an empirical study on token pruning for historical screenshots in GUI scenarios and distill three practical insights that are crucial for designing effective pruning strategies. First, we observe that GUI screenshots exhibit a distinctive foreground-background semantic composition. To probe this property, we apply a simple edge-based separation to partition screenshots into foreground and background regions. Surprisingly, we find that, contrary to the common assumption that background areas have little semantic value, they effectively capture interface-state transitions, thereby providing auxiliary cues for GUI reasoning. Second, compared with carefully designed pruning strategies, random pruning possesses an inherent advantage in preserving spatial structure, enabling better performance under the same computational budget. Finally, we observe that GUI Agents exhibit a recency effect similar to human cognition: by allocating larger token budgets to more recent screenshots and heavily compressing distant ones, we can significantly reduce computational cost while maintaining nearly unchanged performance. These findings offer new insights and practical guidance for the design of efficient GUI visual agents.

Leiyu Wang, Biao Jin, Feng Huang et al.

Oriented object detection for multi-spectral imagery faces significant challenges due to differences both within and between modalities. Although existing methods have improved detection accuracy through complex network architectures, their high computational complexity and memory consumption severely restrict their performance. Motivated by the success of large kernel convolutions in remote sensing, we propose MO R-CNN, a lightweight framework for multi-spectral oriented detection featuring heterogeneous feature extraction network (HFEN), single modality supervision (SMS), and condition-based multimodal label fusion (CMLF). HFEN leverages inter-modal differences to adaptively align, merge, and enhance multi-modal features. SMS constrains multi-scale features and enables the model to learn from multiple modalities. CMLF fuses multimodal labels based on specific rules, providing the model with a more robust and consistent supervisory signal. Experiments on the DroneVehicle, VEDAI and OGSOD datasets prove the superiority of our method. The source code is available at:https://github.com/Iwill-github/MORCNN.

Xuyang Liu, Siteng Huang, Yachen Kang et al.

Large-scale text-to-image diffusion models have shown impressive capabilities for generative tasks by leveraging strong vision-language alignment from pre-training. However, most vision-language discriminative tasks require extensive fine-tuning on carefully-labeled datasets to acquire such alignment, with great cost in time and computing resources. In this work, we explore directly applying a pre-trained generative diffusion model to the challenging discriminative task of visual grounding without any fine-tuning and additional training dataset. Specifically, we propose VGDiffZero, a simple yet effective zero-shot visual grounding framework based on text-to-image diffusion models. We also design a comprehensive region-scoring method considering both global and local contexts of each isolated proposal. Extensive experiments on RefCOCO, RefCOCO+, and RefCOCOg show that VGDiffZero achieves strong performance on zero-shot visual grounding. Our code is available at https://github.com/xuyang-liu16/VGDiffZero.

Jian Jia, Yipei Wang, Yan Li et al.

Contemporary recommendation systems predominantly rely on ID embedding to capture latent associations among users and items. However, this approach overlooks the wealth of semantic information embedded within textual descriptions of items, leading to suboptimal performance and poor generalizations. Leveraging the capability of large language models to comprehend and reason about textual content presents a promising avenue for advancing recommendation systems. To achieve this, we propose an Llm-driven knowlEdge Adaptive RecommeNdation (LEARN) framework that synergizes open-world knowledge with collaborative knowledge. We address computational complexity concerns by utilizing pretrained LLMs as item encoders and freezing LLM parameters to avoid catastrophic forgetting and preserve open-world knowledge. To bridge the gap between the open-world and collaborative domains, we design a twin-tower structure supervised by the recommendation task and tailored for practical industrial application. Through experiments on the real large-scale industrial dataset and online A/B tests, we demonstrate the efficacy of our approach in industry application. We also achieve state-of-the-art performance on six Amazon Review datasets to verify the superiority of our method.

Xuyang Liu, Zichen Wen, Shaobo Wang et al.

The advancement of large language models (LLMs) and multi-modal LLMs (MLLMs) has historically relied on scaling model parameters. However, as hardware limits constrain further model growth, the primary computational bottleneck has shifted to the quadratic cost of self-attention over increasingly long sequences by ultra-long text contexts, high-resolution images, and extended videos. In this position paper, \textbf{we argue that the focus of research for efficient artificial intelligence (AI) is shifting from model-centric compression to data-centric compression}. We position data-centric compression as the emerging paradigm, which improves AI efficiency by directly compressing the volume of data processed during model training or inference. To formalize this shift, we establish a unified framework for existing efficiency strategies and demonstrate why it constitutes a crucial paradigm change for long-context AI. We then systematically review the landscape of data-centric compression methods, analyzing their benefits across diverse scenarios. Finally, we outline key challenges and promising future research directions. Our work aims to provide a novel perspective on AI efficiency, synthesize existing efforts, and catalyze innovation to address the challenges posed by ever-increasing context lengths.

Xinying Lin, Xuyang Liu, Hong Yang et al.

With the advent of image super-resolution (SR) algorithms, how to evaluate the quality of generated SR images has become an urgent task. Although full-reference methods perform well in SR image quality assessment (SR-IQA), their reliance on high-resolution (HR) images limits their practical applicability. Leveraging available reconstruction information as much as possible for SR-IQA, such as low-resolution (LR) images and the scale factors, is a promising way to enhance assessment performance for SR-IQA without HR for reference. In this letter, we attempt to evaluate the perceptual quality and reconstruction fidelity of SR images considering LR images and scale factors. Specifically, we propose a novel dual-branch reduced-reference SR-IQA network, \ie, Perception- and Fidelity-aware SR-IQA (PFIQA). The perception-aware branch evaluates the perceptual quality of SR images by leveraging the merits of global modeling of Vision Transformer (ViT) and local relation of ResNet, and incorporating the scale factor to enable comprehensive visual perception. Meanwhile, the fidelity-aware branch assesses the reconstruction fidelity between LR and SR images through their visual perception. The combination of the two branches substantially aligns with the human visual system, enabling a comprehensive SR image evaluation. Experimental results indicate that our PFIQA outperforms current state-of-the-art models across three widely-used SR-IQA benchmarks. Notably, PFIQA excels in assessing the quality of real-world SR images.

Wenbin Guan, Zijiu Yang, Xiaohong Wu et al.

Presently, the task of few-shot object detection (FSOD) in remote sensing images (RSIs) has become a focal point of attention. Numerous few-shot detectors, particularly those based on two-stage detectors, face challenges when dealing with the multiscale complexities inherent in RSIs. Moreover, these detectors present impractical characteristics in real-world applications, mainly due to their unwieldy model parameters when handling large amount of data. In contrast, we recognize the advantages of one-stage detectors, including high detection speed and a global receptive field. Consequently, we choose the YOLOv7 one-stage detector as a baseline and subject it to a novel meta-learning training framework. This transformation allows the detector to adeptly address FSOD tasks while capitalizing on its inherent advantage of lightweight. Additionally, we thoroughly investigate the samples generated by the meta-learning strategy and introduce a novel meta-sampling approach to retain samples produced by our designed meta-detection head. Coupled with our devised meta-cross loss, we deliberately utilize "negative samples" that are often overlooked to extract valuable knowledge from them. This approach serves to enhance detection accuracy and efficiently refine the overall meta-learning strategy. To validate the effectiveness of our proposed detector, we conducted performance comparisons with current state-of-the-art detectors using the DIOR and NWPU VHR-10.v2 datasets, yielding satisfactory results.

Junjie Chen, Xuyang Liu, Zichen Wen et al.

Large vision-language models (LVLMs) have demonstrated remarkable capabilities in multimodal understanding tasks. However, the increasing demand for high-resolution image and long-video understanding results in substantial token counts, leading to reduced inference efficiency. Token compression offers a direct solution by reducing the number of tokens to be processed, thereby improving computational efficiency. Through extensive analysis, we identify two critical limitations in existing inner-LLM token compression methods: positional bias and incompatibility with efficient operators, which hinder their practical deployment for LVLM acceleration. This paper presents the first approach from a token variation perspective, revealing that visual token variations within LLMs exhibit task-agnostic properties. We propose Variation-aware Vision Token Dropping (\textit{i.e.}, \textbf{V$^2$Drop}), which progressively removes visual tokens with minimal variation during LVLM inference, thereby enhancing computational efficiency. Extensive experiments across multiple models and benchmarks demonstrate that our V$^2$Drop is able to maintain \textbf{94.0\%} and \textbf{98.6\%} of the original model performance for image and video understanding tasks respectively, while reducing LLM generation latency by \textbf{31.5\%} and \textbf{74.2\%}. When combined with efficient operators, V$^2$Drop further reduces GPU peak memory usage.

Honggang Chen, Xiaohai He, Linbo Qing et al.

Single image super-resolution (SISR), which aims to reconstruct a high-resolution (HR) image from a low-resolution (LR) observation, has been an active research topic in the area of image processing in recent decades. Particularly, deep learning-based super-resolution (SR) approaches have drawn much attention and have greatly improved the reconstruction performance on synthetic data. Recent studies show that simulation results on synthetic data usually overestimate the capacity to super-resolve real-world images. In this context, more and more researchers devote themselves to develop SR approaches for realistic images. This article aims to make a comprehensive review on real-world single image super-resolution (RSISR). More specifically, this review covers the critical publically available datasets and assessment metrics for RSISR, and four major categories of RSISR methods, namely the degradation modeling-based RSISR, image pairs-based RSISR, domain translation-based RSISR, and self-learning-based RSISR. Comparisons are also made among representative RSISR methods on benchmark datasets, in terms of both reconstruction quality and computational efficiency. Besides, we discuss challenges and promising research topics on RSISR.

Junxi Feng, Xiaohai He, Qizhi Teng et al.

Porous media are ubiquitous in both nature and engineering applications, thus their modelling and understanding is of vital importance. In contrast to direct acquisition of three-dimensional (3D) images of such medium, obtaining its sub-region (s) like two-dimensional (2D) images or several small areas could be much feasible. Therefore, reconstructing whole images from the limited information is a primary technique in such cases. Specially, in practice the given data cannot generally be determined by users and may be incomplete or partially informed, thus making existing reconstruction methods inaccurate or even ineffective. To overcome this shortcoming, in this study we proposed a deep learning-based framework for reconstructing full image from its much smaller sub-area(s). Particularly, conditional generative adversarial network (CGAN) is utilized to learn the mapping between input (partial image) and output (full image). To preserve the reconstruction accuracy, two simple but effective objective functions are proposed and then coupled with the other two functions to jointly constrain the training procedure. Due to the inherent essence of this ill-posed problem, a Gaussian noise is introduced for producing reconstruction diversity, thus allowing for providing multiple candidate outputs. Extensively tested on a variety of porous materials and demonstrated by both visual inspection and quantitative comparison, the method is shown to be accurate, stable yet fast ($\sim0.08s$ for a $128 \times 128$ image reconstruction). We highlight that the proposed approach can be readily extended, such as incorporating any user-define conditional data and an arbitrary number of object functions into reconstruction, and being coupled with other reconstruction methods.

Honggang Chen, Xiaohai He, Linbo Qing et al.

JPEG is one of the widely used lossy compression methods. JPEG-compressed images usually suffer from compression artifacts including blocking and blurring, especially at low bit-rates. Soft decoding is an effective solution to improve the quality of compressed images without changing codec or introducing extra coding bits. Inspired by the excellent performance of the deep convolutional neural networks (CNNs) on both low-level and high-level computer vision problems, we develop a dual pixel-wavelet domain deep CNNs-based soft decoding network for JPEG-compressed images, namely DPW-SDNet. The pixel domain deep network takes the four downsampled versions of the compressed image to form a 4-channel input and outputs a pixel domain prediction, while the wavelet domain deep network uses the 1-level discrete wavelet transformation (DWT) coefficients to form a 4-channel input to produce a DWT domain prediction. The pixel domain and wavelet domain estimates are combined to generate the final soft decoded result. Experimental results demonstrate the superiority of the proposed DPW-SDNet over several state-of-the-art compression artifacts reduction algorithms.

Honggang Chen, Xiaohai He, Chao Ren et al.

In recent years, much research has been conducted on image super-resolution (SR). To the best of our knowledge, however, few SR methods were concerned with compressed images. The SR of compressed images is a challenging task due to the complicated compression artifacts, while many images suffer from them in practice. The intuitive solution for this difficult task is to decouple it into two sequential but independent subproblems, i.e., compression artifacts reduction (CAR) and SR. Nevertheless, some useful details may be removed in CAR stage, which is contrary to the goal of SR and makes the SR stage more challenging. In this paper, an end-to-end trainable deep convolutional neural network is designed to perform SR on compressed images (CISRDCNN), which reduces compression artifacts and improves image resolution jointly. Experiments on compressed images produced by JPEG (we take the JPEG as an example in this paper) demonstrate that the proposed CISRDCNN yields state-of-the-art SR performance on commonly used test images and imagesets. The results of CISRDCNN on real low quality web images are also very impressive, with obvious quality enhancement. Further, we explore the application of the proposed SR method in low bit-rate image coding, leading to better rate-distortion performance than JPEG.