Recent transformer-based methods for SISR reduce computational costs by restricting self-attention to fixed windows. However, rectangular window cuts can be rigid and prone to non-similar structure interference and edge artifacts. Superpixel-based self-attention offers more flexibility and better captures local interactions, but its boundaries remain prone to distortion. To address these limitations, we propose a High-Frequency Superpixel Interaction Network (HFSIN). Superpixels generally contain low-frequency information internally, while their edges correspond to high-frequency details. We propose the High-Frequency Superpixel Attention (HFSA), which combines high-frequency priors with superpixel token interactions, enabling more accurate preservation of sharp edges and fine details. Additionally, our Superpixel Intra-Attention (SIA) improves internal feature aggregation. HFSIN activates superpixel tokens across more image regions, capturing richer features and improving SR performance. Experiments demonstrate that HFSIN outperforms state-of-the-art methods on several benchmark datasets, both quantitatively and visually.

Small object detection poses significant challenges, including the inability to extract effective features from diminutive targets, interference from complex background noise, and the loss of weak features in deep networks. Additionally, the Intersection over Union (IoU) metric is overly sensitive to minor offsets between predicted and ground truth bounding boxes for small objects. To address these issues, we propose a novel detection framework, termed Multi-Scale Contextual Robust Optimization-YOLO (MCRO-YOLO). First, we introduce a new feature extraction module called Cross Stage Partial with Dual-Scale Fusion (C2DF), designed to suppress background noise while balancing the extraction of local details and global contextual information for robust feature representation. Second, we propose the Dual-Scale Adaptive Weighted Attention Relay Skip Network (DWS-RelayNet), which reconstructs and enhances small object features during deep network propagation. Finally, we present a novel detection loss metric, Flexi-IoU (FIoU), which effectively mitigates the excessive sensitivity of small object loss values to minor positional offsets. Extensive experiments on the VisDrone dataset demonstrate that MCRO-YOLO surpasses state-of-the-art methods, achieving superior small object detection accuracy while maintaining computational efficiency.

Although interactive image segmentation techniques have made significant progress, supervised learning-based methods rely heavily on large-scale labeled data which is difficult to obtain in certain domains such as medicine, biology, etc. Models trained on natural images also struggle to achieve satisfactory results when directly applied to these domains. To solve this dilemma, we propose a Self-supervised Interactive Segmentation (SIS) method that achieves superior generalization performance. By clustering features from unlabeled data, we obtain classifiers that assign pseudo-labels to pixels in images. After refinement by super-pixel voting, these pseudo-labels are then used to train our segmentation network. To enable our network to better adapt to cross-domain images, we introduce correction learning and anti-forgetting regularization to conduct test-time adaptation. Our experiment results on five datasets show that our approach significantly outperforms other interactive segmentation methods across natural image datasets in the same conditions and achieves even better performance than some supervised methods when across to medical image domain. The code and models are available at https://github.com/leal0110/SIS.

We address the problem of full body human pose estimation in video. Most previous work consider body part, pose or trajectory of body part as basic unit to compose the pose sequence. In contrast, we consider tracklet of body part as the basic unit. Based on this medium granularity representation we develop a spatio-temporal graphical model to select an optimal tracklet for each part in each video segment. In our model, tracklet nodes of symmetric parts are coupled to one node to overcome the double counting problem. Through iterative spatial and temporal parsing, optimal solution is achieved in polynomial time. We apply our model on three publicly available datasets and show remarkable quantitative and qualitative improvements over the state-of-the-art approaches.

Our goal is to estimate contour flow (the contour pairs with consistent point correspondence) from inconsistent contours extracted independently in two video frames. We formulate the contour flow estimation locally as a motion segmentation problem where motion patterns grouped from optical flow field are exploited for local correspondence measurement. To solve local ambiguities, contour flow estimation is further formulated globally as a contour alignment problem. We propose a novel two-staged strategy to obtain global consistent point correspondence under various contour transitions such as splitting, merging and branching. The goal of the first stage is to obtain possible accurate contour-to-contour alignments, and the second stage aims to make a consistent fusion of many partial alignments. Such a strategy can properly balance the accuracy and the consistency, which enables a middle-level motion representation to be constructed by just concatenating frame-by-frame contour flow estimation. Experiments prove the effectiveness of our method.

We present a novel method to estimate full-body human pose in video sequence by incorporating global motion cues. It has been demonstrated that temporal constraints can largely enhance the pose estimation. Most current approaches typically employ local motion to propagate pose detections to supplement the pose candidates. However, the local motion estimation is often inaccurate under fast movements of body parts and unhelpful when no strong detections achieved in adjacent frames. In this paper, we propose to propagate the detection in each frame using the global motion estimation. Benefiting from the strong detections, our algorithm first produces reasonable trajectory hypotheses for each body part. Then, we cast pose estimation as an optimization problem defined on these trajectories with spatial links between body parts. In the optimization process, we select body part trajectory rather than body part candidate to infer the human pose. Experimental results demonstrate significant performance improvement in comparison with the state-of-the-art methods.