Image composition involves seamlessly integrating given objects into a specific visual context. Current training-free methods rely on composing attention weights from several samplers to guide the generator. However, since these weights are derived from disparate contexts, their combination leads to coherence confusion and loss of appearance information. These issues worsen with their excessive focus on background generation, even when unnecessary in this task. This not only impedes their swift implementation but also compromises foreground generation quality. Moreover, these methods introduce unwanted artifacts in the transition area. In this paper, we formulate image composition as a subject-based local editing task, solely focusing on foreground generation. At each step, the edited foreground is combined with the noisy background to maintain scene consistency. To address the remaining issues, we propose PrimeComposer, a faster training-free diffuser that composites the images by well-designed attention steering across different noise levels. This steering is predominantly achieved by our Correlation Diffuser, utilizing its self-attention layers at each step. Within these layers, the synthesized subject interacts with both the referenced object and background, capturing intricate details and coherent relationships. This prior information is encoded into the attention weights, which are then integrated into the self-attention layers of the generator to guide the synthesis process. Besides, we introduce a Region-constrained Cross-Attention to confine the impact of specific subject-related tokens to desired regions, addressing the unwanted artifacts shown in the prior method thereby further improving the coherence in the transition area. Our method exhibits the fastest inference efficiency and extensive experiments demonstrate our superiority both qualitatively and quantitatively. The code is available at https://github.com/CodeGoat24/PrimeComposer.

Tensor Nuclear Norm (TNN) has been widely employed as a regularization term for hyperspectral image super-resolution (HSISR). However, conventional TNN constraints based on Discrete Fourier Transform (DFT) often suffer from rank estimation biases and an inability to effectively capture complex spectral-spatial correlations, limiting their efficacy in HSISR. To address these challenges, we propose a Nonlinear Learnable Triple-domain (NLT) transform framework that integrates nonlinear transform, DFT, and self-learning adaptation. This multi-stage process promotes singular value concentration, improving low-rank approximation and rank estimation accuracy. Building upon this framework, we develop an NL-transform-oriented tensor product, a truncated singular value decomposition (TSVD) operation, and a novel tensor nuclear norm (NLTN) tailored for HSISR. By incorporating spectral subspace estimation and clustering-based patch grouping, our approach effectively leverages spatial-spectral correlations and non-local self-similarities, leading to enhanced reconstruction quality. To further mitigate singular value over-penalization, we introduce a logarithmic-based generalized NLTNN (GNLTN) and formulate an optimization strategy based on the alternating direction method of multipliers (ADMM). Extensive experiments demonstrate that our method significantly outperforms existing approaches in terms of fusion accuracy and visual fidelity, setting new benchmarks for hyperspectral image super-resolution. The code is available at https://github.com/xuhonghui96/GNLTN.

Changhai Zhou, Yuhua Zhou, Yibin Wang, Shijie Han, Qian Qiao, Hongguang Li. Findings of the Association for Computational Linguistics: NAACL 2025. 2025.