In this paper, we present Diffusion-4K, a novel framework for direct ultra-high-resolution image synthesis using text-to-image diffusion models. The core advancements include: (1) Aesthetic-4K Benchmark: addressing the absence of a publicly available 4K image synthesis dataset, we construct Aesthetic-4K, a comprehensive benchmark for ultra-high-resolution image generation. We curated a high-quality 4K dataset with carefully selected images and captions generated by GPT-4o. Additionally, we introduce GLCM Score and Compression Ratio metrics to evaluate fine details, combined with holistic measures such as FID, Aesthetics and CLIPScore for a comprehensive assessment of ultra-high-resolution images. (2) Wavelet-based Fine-tuning: we propose a wavelet-based fine-tuning approach for direct training with photorealistic 4K images, applicable to various latent diffusion models, demonstrating its effectiveness in synthesizing highly detailed 4K images. Consequently, Diffusion-4K achieves impressive performance in high-quality image synthesis and text prompt adherence, especially when powered by modern large-scale diffusion models (e.g., SD3-2B and Flux-12B). Extensive experimental results from our benchmark demonstrate the superiority of Diffusion-4K in ultra-high-resolution image synthesis. Code is available at https://github.com/zhang0jhon/diffusion-4k.

Due to the irregularity of 3D point clouds, it is extremely challenging to detect the most salient objects from them and segment contours. DSFormer is proposed for 3D salient object detection, addressing challenges like small objects, multiple objects and complex backgrounds. It employs an encoder-decoder structure with multi-scale transformers and deformable convolutions to enhance modeling flexibility. The SPB module (Saliency Perception Block) in the decoder boosts high-level features. DSFormer shows notable improvements on both indoor and outdoor saliency datasets and is validated through ablation studies and visualizations.

Recently, MLP-based architectures have achieved competitive performance with convolutional neural networks (CNNs) and vision transformers (ViTs) across various vision tasks. However, most MLP-based methods introduce local feature interactions to facilitate direct adaptation to downstream tasks, thereby lacking the ability to capture global visual dependencies and multi-scale context, ultimately resulting in unsatisfactory performance on dense prediction. This paper proposes a competitive and effective MLP-based architecture called Pyramid Fusion MLP (PFMLP) to address the above limitation. Specifically, each block in PFMLP introduces multi-scale pooling and fully connected layers to generate feature pyramids, which are subsequently fused using up-sample layers and an additional fully connected layer. Employing different down-sample rates allows us to obtain diverse receptive fields, enabling the model to simultaneously capture long-range dependencies and fine-grained cues, thereby exploiting the potential of global context information and enhancing the spatial representation power of the model. Our PFMLP is the first lightweight MLP to obtain comparable results with state-of-the-art CNNs and ViTs on the ImageNet-1K benchmark.With larger FLOPs, it exceeds state-of-the-art CNNs, ViTs, and MLPs under similar computational complexity. Furthermore, experiments in object detection, instance segmentation, and semantic segmentation demonstrate that the visual representation acquired from PFMLP can be seamlessly transferred to downstream tasks, producing competitive results. All materials contain the training codes and logs are released at https://github.com/huangqiuyu/PFMLP.

In semi-supervised learning (SSL), many approaches follow the effective self-training paradigm with consistency regularization, utilizing threshold heuristics to alleviate label noise. However, such threshold heuristics lead to the underutilization of crucial discriminative information from the excluded data. In this paper, we present OTAMatch, a novel SSL framework that reformulates pseudo-labeling as an optimal transport (OT) assignment problem and simultaneously exploits data with high confidence to mitigate the confirmation bias. Firstly, OTAMatch models the pseudo-label allocation task as a convex minimization problem, facilitating end-to-end optimization with all pseudo-labels and employing the Sinkhorn-Knopp algorithm for efficient approximation. Meanwhile, we incorporate epsilon-greedy posterior regularization and curriculum bias correction strategies to constrain the distribution of OT assignments, improving the robustness with noisy pseudo-labels. Secondly, we propose PseudoNCE, which explicitly exploits pseudo-label consistency with threshold heuristics to maximize mutual information within self-training, significantly boosting the balance of convergence speed and performance. Consequently, our proposed approach achieves competitive performance on various SSL benchmarks. Specifically, OTAMatch substantially outperforms the previous state-of-the-art SSL algorithms in realistic and challenging scenarios, exemplified by a notable 9.45% error rate reduction over SoftMatch on ImageNet with 100K-label split, underlining its robustness and effectiveness.

Detecting salient objects in videos is a very challenging task. Current state-of-the-art methods are dominated by motion based deep neural networks, among which optical flow is often leveraged as motion representation. Though with robust performance, these optical flow-based video salient object detection methods face at least two problems that may hinder their generalization and application. First, computing optical flow as a pre-processing step does not support direct end-to-end learning; second, little attention has been given to the quality of visual features due to high computational cost of spatiotemporal feature encoding. In this paper we propose a novel self-sufficient feature enhancing network (SFENet) for video salient object detection, which leverages optical flow estimation as an auxiliary task while being end-to-end trainable. With a joint training scheme of both salient object detection and optical flow estimation, its multi-task architecture can be totally self-sufficient for achieving good performance without any pre-processing. Furthermore, for improving feature quality, we design four lightweight modules in spatial and temporal domains, including cross-layer fusion, multi-level warping, spatial-channel attention and boundary-aware refinement. The proposed method is evaluated through extensive experiments on five video salient object detection datasets. Experimental results show that our SFENet can be easily trained with fast convergence speed. It significantly outperforms previous methods in terms of various evaluation metrics. Moreover, with optical flow estimation and unsupervised video object segmentation as example applications, our method also yields state-of-the-art results on standard datasets.

This work addresses the task of action recognition in video sequences. In real world applications, this task is quite challenging due to the complex background of video content, the similarities between different types of actions, the dependence on a large amount of annotated data, and so on. Most of the existing methods fail to distinguish similar actions with the same static appearance and motion pattern. We attempt to address this issue from the perspective of a local-global view, considering videos as combinations of a set of action units (local semantic information) and their relations along temporal dimension (global relation information). To achieve this end, we propose a novel Local-global Networks (LgNet) to enhance recognition of similar action. Besides, we propose an end-to-end training method to decrease the reliance on annotated data. It combines self-supervised learning and supervised learning, which not only enables the model to learn video representations from a large number unannotated data but also avoids subsequent finetuning. The proposed training method can be flexibly equipped to a wide array of vision tasks. Experiments on several benchmark datasets show that our proposed model and training method achieve state-of-the-art performance.