arXiv subjects

Zehong Ma

Publications and source records attributed to Zehong Ma.

6 recordsLinked to original sources

arXiv2026

PixelGen: Improving Pixel Diffusion with Perceptual Supervision

Pixel diffusion generates images directly in pixel space, avoiding the VAE artifacts and representational bottlenecks of two-stage latent diffusion. Recent JiT further simplifies pixel diffusion with x-prediction, where the model predicts clean images rather than velocity. However, the standard pixel-wise diffusion loss treats all pixels equally, spending model capacity to perceptually insignificant signals and often leading to blurry samples. We propose PixelGen, an end-to-end pixel diffusion framework that augments x-prediction with perceptual supervision. Specifically, PixelGen introduces two complementary perceptual losses on top of x-prediction: an LPIPS loss for local textures and a P-DINO loss for global semantics. To preserve sample coverage, PixelGen further proposes a noise-gating strategy that applies these losses only at lower-noise timesteps. On ImageNet-256 without classifier-free guidance, PixelGen achieves an FID of 5.11 in 80 training epochs, surpassing the latent diffusion baselines. Moreover, PixelGen scales efficiently to text-to-image generation, reaching a GenEval score of 0.79 with only 6 days of training on 8xH800 GPUs. These results show that perceptual supervision substantially narrows the gap between pixel and latent diffusion while preserving a simple one-stage pipeline. Codes are available at https://github.com/Zehong-Ma/PixelGen.

cs.CV↗

arXiv2025

DeCo: Frequency-Decoupled Pixel Diffusion for End-to-End Image Generation

Pixel diffusion aims to generate images directly in pixel space in an end-to-end fashion. This approach avoids the limitations of VAE in the two-stage latent diffusion, offering higher model capacity. Existing pixel diffusion models suffer from slow training and inference, as they usually model both high-frequency signals and low-frequency semantics within a single diffusion transformer (DiT). To pursue a more efficient pixel diffusion paradigm, we propose the frequency-DeCoupled pixel diffusion framework. With the intuition to decouple the generation of high and low frequency components, we leverage a lightweight pixel decoder to generate high-frequency details conditioned on semantic guidance from the DiT. This thus frees the DiT to specialize in modeling low-frequency semantics. In addition, we introduce a frequency-aware flow-matching loss that emphasizes visually salient frequencies while suppressing insignificant ones. Extensive experiments show that DeCo achieves superior performance among pixel diffusion models, attaining FID of 1.62 (256x256) and 2.22 (512x512) on ImageNet, closing the gap with latent diffusion methods. Furthermore, our pretrained text-to-image model achieves a leading overall score of 0.86 on GenEval in system-level comparison. Codes are publicly available at https://github.com/Zehong-Ma/DeCo.

cs.CV↗

arXiv2025

MagCache: Fast Video Generation with Magnitude-Aware Cache

Existing acceleration techniques for video diffusion models often rely on uniform heuristics or time-embedding variants to skip timesteps and reuse cached features. These approaches typically require extensive calibration with curated prompts and risk inconsistent outputs due to prompt-specific overfitting. In this paper, we introduce a novel and robust discovery: a unified magnitude law observed across different models and prompts. Specifically, the magnitude ratio of successive residual outputs decreases monotonically, steadily in most timesteps while rapidly in the last several steps. Leveraging this insight, we introduce a Magnitude-aware Cache (MagCache) that adaptively skips unimportant timesteps using an error modeling mechanism and adaptive caching strategy. Unlike existing methods requiring dozens of curated samples for calibration, MagCache only requires a single sample for calibration. Experimental results show that MagCache achieves 2.10x-2.68x speedups on Open-Sora, CogVideoX, Wan 2.1, and HunyuanVideo, while preserving superior visual fidelity. It significantly outperforms existing methods in LPIPS, SSIM, and PSNR, under similar computational budgets.

cs.CV↗

arXiv2025

Efficient Multi-modal Long Context Learning for Training-free Adaptation

Traditional approaches to adapting multi-modal large language models (MLLMs) to new tasks have relied heavily on fine-tuning. This paper introduces Efficient Multi-Modal Long Context Learning (EMLoC), a novel training-free alternative that embeds demonstration examples directly into the model input. EMLoC offers a more efficient, flexible, and scalable solution for task adaptation. Because extremely lengthy inputs introduce prohibitive computational and memory overhead, EMLoC contributes a chunk-wise compression mechanism combined with layer-wise adaptive pruning. It condenses long-context multimodal inputs into compact, task-specific memory representations. By adaptively pruning tokens at each layer under a Jensen-Shannon divergence constraint, our method achieves a dramatic reduction in inference complexity without sacrificing performance. This approach is the first to seamlessly integrate compression and pruning techniques for multi-modal long-context learning, offering a scalable and efficient solution for real-world applications. Extensive experiments on diverse vision-language benchmarks demonstrate that EMLoC achieves performance on par with or superior to naive long-context approaches. Our results highlight the potential of EMLoC as a groundbreaking framework for efficient and flexible adaptation of multi-modal models in resource-constrained environments. Codes are publicly available at https://github.com/Zehong-Ma/EMLoC.

cs.CV↗

arXiv2025

Multi-modal Reference Learning for Fine-grained Text-to-Image Retrieval

Fine-grained text-to-image retrieval aims to retrieve a fine-grained target image with a given text query. Existing methods typically assume that each training image is accurately depicted by its textual descriptions. However, textual descriptions can be ambiguous and fail to depict discriminative visual details in images, leading to inaccurate representation learning. To alleviate the effects of text ambiguity, we propose a Multi-Modal Reference learning framework to learn robust representations. We first propose a multi-modal reference construction module to aggregate all visual and textual details of the same object into a comprehensive multi-modal reference. The multi-modal reference hence facilitates the subsequent representation learning and retrieval similarity computation. Specifically, a reference-guided representation learning module is proposed to use multi-modal references to learn more accurate visual and textual representations. Additionally, we introduce a reference-based refinement method that employs the object references to compute a reference-based similarity that refines the initial retrieval results. Extensive experiments are conducted on five fine-grained text-to-image retrieval datasets for different text-to-image retrieval tasks. The proposed method has achieved superior performance over state-of-the-art methods. For instance, on the text-to-person image retrieval dataset RSTPReid, our method achieves the Rank1 accuracy of 56.2\%, surpassing the recent CFine by 5.6\%.

cs.CV↗

arXiv2024

OVMR: Open-Vocabulary Recognition with Multi-Modal References

The challenge of open-vocabulary recognition lies in the model has no clue of new categories it is applied to. Existing works have proposed different methods to embed category cues into the model, \eg, through few-shot fine-tuning, providing category names or textual descriptions to Vision-Language Models. Fine-tuning is time-consuming and degrades the generalization capability. Textual descriptions could be ambiguous and fail to depict visual details. This paper tackles open-vocabulary recognition from a different perspective by referring to multi-modal clues composed of textual descriptions and exemplar images. Our method, named OVMR, adopts two innovative components to pursue a more robust category cues embedding. A multi-modal classifier is first generated by dynamically complementing textual descriptions with image exemplars. A preference-based refinement module is hence applied to fuse uni-modal and multi-modal classifiers, with the aim to alleviate issues of low-quality exemplar images or textual descriptions. The proposed OVMR is a plug-and-play module, and works well with exemplar images randomly crawled from the Internet. Extensive experiments have demonstrated the promising performance of OVMR, \eg, it outperforms existing methods across various scenarios and setups. Codes are publicly available at \href{https://github.com/Zehong-Ma/OVMR}{https://github.com/Zehong-Ma/OVMR}.

cs.CV↗