Search

The contrastive vision–language pre-trained model CLIP, driven by large-scale open-vocabulary image–text pairs, has recently demonstrated remarkable zero-shot generalization capabilities in diverse downstream image tasks, which has made numerous models dominated by the "image pre-training followed by fine-tuning" paradigm exhibit promising results on standard video benchmarks. However, as models scale up, full fine-tuning adaptive strategy for specific tasks becomes difficult in terms of training and storage. In this work, we propose a novel method that adapts CLIP to the video domain for efficient recognition without destroying the original pre-trained parameters. Specifically, we introduce temporal prompts to realize the object of reasoning about the dynamic content of videos for pre-trained models that lack temporal cues. Then, by replacing the direct learning style of prompt vectors with a lightweight reparameterization encoder, the model can be adapted to domain-specific adjustment to learn more generalizable representations. Furthermore, we predefine a Chinese label dictionary to enhance video representation by co-supervision of Chinese and English semantics. Extensive experiments on video action recognition benchmarks show that our method achieves competitive or even better performance than most existing methods with fewer trainable parameters in both general and few-shot recognition scenarios. [ABSTRACT FROM AUTHOR]

In this paper, we present an efficient method for constructing symmetric subdivision schemes reproducing high-degree polynomials, without solving a system of linear equations. Original symmetric subdivision and its deduced subdivisions have similar increasing characteristics to the family of pseudo-splines from the polynomial reproduction point of view. Several examples are given to illustrate the efficiency of the method. [ABSTRACT FROM AUTHOR]

For non-uniform control polygons, a parameterized four-point interpolation curve ternary subdivision scheme is proposed, and its convergence and continuity are demonstrated. Following curve subdivision, a non-uniform interpolation surface ternary subdivision on regular quadrilateral meshes is proposed by applying the tensor product method. Analyses were conducted on the updating rules of parameters, proving that the limit surface is continuous. In this paper, we present a novel interpolation subdivision method to generate new virtual edge points and new face points of the extraordinary points of quadrilateral mesh. We also provide numerical examples to assess the validity of various interpolation methods. [ABSTRACT FROM AUTHOR]

Mesh denoising is a classical task in mesh processing. Many state-of-the-art methods are still unable to quickly and robustly denoise multifarious noisy 3D meshes, especially in the case of high noise. Recently, neural network-based models have played a leading role in natural language, audio, image, video, and 3D model processing. Inspired by these works, we propose a data-driven mesh denoising method based on recurrent neural networks, which learns the relationship between the feature descriptors and the ground-truth normals. The recurrent neural network has a feedback loop before entering the output layer. By means of the self-feedback of neurons, the output of a recurrent neural network is related not only to the current input but also to the output of the previous moments. To deal with meshes with various geometric features, we use k-means to cluster the faces of the mesh according to geometric similarity and train neural networks for each category individually in the offline learning stage. Each network model, acting similar to a normal regression function, will map the geometric feature descriptor of each facet extracted from the mesh to the denoised facet normal. Then, the denoised normals are used to calculate the new feature descriptors, which become the input of the next similar regression model. In this system, three normal regression modules are cascaded to generate the last facet normals. Lastly, the model's vertex positions are updated according to the denoised normals. A large number of visual and numerical results have demonstrated that the proposed model outperforms the state-of-the-art methods in most cases. [ABSTRACT FROM AUTHOR]

In many application fields (closed curve noise data reconstruction, time series data fitting, image edge smoothing, skeleton extraction, etc.), curve reconstruction based on noise data has always been a popular but challenging problem. In a single domain, there are many methods for curve reconstruction of noise data, but a method suitable for multi-domain curve reconstruction has received much less attention in the literature. More importantly, the existing methods have shortcomings in time consumption when dealing with large data and high-density point cloud curve reconstruction. For this reason, we hope to propose a curve fitting algorithm suitable for many fields and low time consumption. In this paper, a curve reconstruction method based on clustering and point cloud principal component analysis is proposed. Firstly, the point cloud is clustered by the K++ means algorithm. Secondly, a denoising method based on point cloud principal component analysis is proposed to obtain the interpolation nodes of curve subdivision. Finally, the fitting curve is obtained by the parametric curve subdivision method. Comparative experiments show that our method is superior to the classical fitting method in terms of time consumption and effect. In addition, our method is not constrained by the shape of the point cloud, and can play a role in time series data, image thinning and edge smoothing. [ABSTRACT FROM AUTHOR]

An enormous number of CNN classification algorithms have been proposed in the literature. Nevertheless, in these algorithms, appropriate filter size selection, data preparation, limitations in datasets, and noise have not been taken into consideration. As a consequence, most of the algorithms have failed to make a noticeable improvement in classification accuracy. To address the shortcomings of these algorithms, our paper presents the following contributions: Firstly, after taking the domain knowledge into consideration, the size of the effective receptive field (ERF) is calculated. Calculating the size of the ERF helps us to select a typical filter size which leads to enhancing the classification accuracy of our CNN. Secondly, unnecessary data leads to misleading results and this, in turn, negatively affects classification accuracy. To guarantee the dataset is free from any redundant or irrelevant variables to the target variable, data preparation is applied before implementing the data classification mission. Thirdly, to decrease the errors of training and validation, and avoid the limitation of datasets, data augmentation has been proposed. Fourthly, to simulate the real-world natural influences that can affect image quality, we propose to add an additive white Gaussian noise with σ = 0.5 to the MNIST dataset. As a result, our CNN algorithm achieves state-of-the-art results in handwritten digit recognition, with a recognition accuracy of 99.98%, and 99.40% with 50% noise. [ABSTRACT FROM AUTHOR]