Capturing sufficient global context and rich spatial structure information is critical for dense prediction tasks. Convolutional Neural Network (CNN) is particularly adept at modeling fine-grained local features, while Transformer excels at modeling global context information. It is evident that CNN and Transformer exhibit complementary characteristics. Exploring the design of a network, that efficiently fuses these two models to leverage their strengths fully and achieve more accurate detection, represents a promising and worthwhile research topic. In this paper, we introduce a novel CNN-Transformer Iterative Fusion Network (CTIF-Net) for salient object detection. It efficiently combines CNN and Transformer to achieve superior performance by using a parallel dual encoder structure and a feature iterative fusion module. Firstly, CTIF-Net extracts features from the image using the CNN and the Transformer, respectively. Secondly, two feature convertors and a feature iterative fusion module are employed to combine and iteratively refine the two sets of features. The experimental results on multiple SOD datasets show that CTIF-Net outperforms 17 state-of-the-art methods, achieving higher performance in various mainstream evaluation metrics such as F-measure, S-measure, and MAE value. The code will be publicly available.

Abstract In clinical medicine, the segmentation of blood vessels in retinal images is essential for subsequent analysis in clinical diagnosis. However, retinal images are often noisy and their vascular structure is relatively tiny, which poses significant challenges for vessel segmentation. To improve the performance of vessel segmentation, an improved model IterNet++ based on the architecture of IterNet is proposed. First, curvelet signal analysis is applied to enhance retinal images. Second, residual convolution (ResConv) blocks and guided filters are introduced to utilise the encoder features of previous iterations in the model to reduce overfitting. Third, offline hard‐sample mining is used to improve segmentation performance by utilising training samples with low segmentation accuracy as many possible on a few‐sample training set. In addition, a test‐time augmentation method is applied to testing samples in test dataset during inference. Extensive experiments show that this model achieves Dice scores of 0.8313, 0.8277, and 0.8372 on DRIVE, CHASE‐DB1, and STARE datasets, respectively, demonstrating the best performance compared with IterNet and other baseline models.

Vertical retargeting for stereoscopic images using seam manipulation-based approaches has remained an open challenge over the years. Even though horizontal retargeting had attracted a huge amount of interest, its seam coupling strategies were not capable to construct valid seam pairs for vertical retargeting. In this article, we propose two seam coupling strategies for vertical retargeting, namely, real mapping and virtual mapping. Our proposed mapping strategies were implemented to address the problems of multiple assignments and missing assignments, which are able to occur in the straightforward generalization from horizontal retargeting to vertical retargeting. On the basis of our proposed method, stereo seams were allowed to lay across occluded regions and occluding regions in stereo images. We maintained the geometric consistency by removing occluded pixels and corresponding occluding pixels in both stereo images. As a result, our method guarantees valid and geometrically consistent stereo seam pairs to be found in the horizontal direction. We generate vertically retargeted stereo images by removing or adding horizontal seam pairs iteratively. We conducted experiments on a number of indoor and outdoor scenes. Experimental results demonstrated that our method overcomes the limitations of vertical retargeting and is effective in preserving the geometric consistency.

To implement vertical retargeting for stereoscopic images, this paper proposes an occlusion-guided stereoscopic image retargeting method via the pixel fusion technique. Traditional seam searching-based methods cannot construct valid horizontal seam pairs due to the existence of occluded regions in stereoscopic images and thus fail to implement vertical retargeting. To solve this issue, we propose a novel horizontal seam coupling strategy guided by the occlusion regions that appear on both sides of stereoscopic images. Horizontal seams were able to be laid across the occluded and occluding regions with their geometric consistency maintained. Another important contribution of our method is incorporating occluding masks into energy optimization. The experimental results show that our method can achieve promising performances in both visual experience and depth preservation.

Retrieving 3D shapes from 2D human sketches has received increasing attention in computer vision and computer graphics. Most previous methods projected 3D shapes from numerous viewpoints and then extracted features of 3D shapes from these projections and calculated the similarity with sketches. However, due to the unknown pose of 3D shapes, viewpoints were usually sampled uniformly from a sphere coordinate. Hence, some projections acquired insufficient descriptions of 3D shapes. In this paper, we proposed a view selection algorithm to find the most reasonable viewpoints, which can benefit representation learning for 3D shapes. Additionally, to indicate the apparent discrepancy between sketches and 3D shapes, we leveraged a generalized similarity model to encourage the accuracy of cross-domain feature matching. We first computed line renderings of 3D shapes from an enormous number of viewpoints. Then, we calculated the similarity of shapes between line renderings and sketches. In this vein, we obtained several superior projections. Second, we implemented a sketch network to extract features of the sketch and a shape network to extract features of projections. We combined the features of different projections to secure the compact representation for 3D shapes. Finally, a metric network was constructed using a cross-domain similarity model, and we trained the metric network with triplet loss. Online hard sample mining was leveraged to accelerate the convergence of the network. We evaluated our method on SHREC‘13 and SHREC’14 sketch track benchmark datasets. The experimental results demonstrated that both view selection and cross-domain similarity models were able to encourage retrieval performance.

Stereo image retargeting plays a significant role in the field of image processing, which aims at making major objects as prominent as possible when the resolution of an image is changed, including maintaining disparity and depth information at the same time. Some seam carving methods are proposed to preserve the geometric consistency of the images. However, the regions of occlusion on both sides are not considered properly. In this article, we propose a solution to solve this problem. A new strategy of seams finding is designed by considering occluded and occluding regions on both of the input images, and leaving geometric consistency in both images intact. We also introduced the method of line segment detection and superpixel segmentation to further improve the quality of the images. Imaging effects are optimized in the process and visual comfort, which is also influenced by other factors, can be boosted as well.

Confusing visual appearance and scattered small-scale patterns commonly exist in natural images, which forms a challenge for prior saliency detection methods. Inspired by the sensitivity to edge information of Human Visual Systems, we propose a universal edge-oriented framework to improve the performance of existing salient detection methods. Firstly, edge probability map is extracted from images and utilized to get edge-based over segmentation. Secondly, merging segments by a hierarchical model to generate edge regions. Finally, the proposed framework turns saliency detection to assign a saliency value to each edge region. Experimental results demonstrate the effectiveness of our framework.

As biomedical data in 3D space collected increasingly, there is a pressing need for efficient and accurate applications in the field of bioinformation analysis. For biomedical purpose, mesh subdivision techniques are commonly used to generate adaptive multi-resolution meshes for fast or accurate algorithms. However, current smoothing methods for each subdivision algorithm will moderate edge and vertex features from the original mesh. In this paper, we propose a feature preserved mesh subdivision framework, which generates a visually sensitive and a more precise result compared with commonly used subdivision methods, to preserve edge and vertex geometrical features of biomedical data.

Drosophila embryonic images provide valuable spatial and temporal information of gene expression. Extraction of the contour of a targeting embryo in an embryonic image is a fundamental step of a computational system for the study of gene-gene interaction on Drosophila. In this paper, we propose a shape model for contour extraction of Drosophila embryos. The shape model is built on connected components of edge pixels. It approximates a connected component of edge pixels by a polygon that can be either convex or concave. The main contribution of the proposed shape model is its ability of segmenting embryos touching each other. Moreover, the proposed shape model is adaptable to a wide range of applications on contour extraction.

Rectangular shape detection has a wide range of applications, such as license plate detection, vehicle detection and building detection. In this paper, we propose a robust framework for rectangular shape detection based on the channel-scale space of RGB images. The framework consists of algorithms developed to address two issues of a shape attention (i.e., a connected component of edge points), including: i) openness and ii) fragmentation. Furthermore, we propose an interestness measure for rectangular shapes by integrating interest points Our case study on license plate detection shows the promise of the proposed framework.