Search

In recent years, remarkable results have been achieved in self-supervised action recognition using skeleton sequences with contrastive learning. It has been observed that the semantic distinction of human action features is often represented by local body parts, such as legs or hands, which are advantageous for skeleton-based action recognition. This paper proposes an attention-based contrastive learning framework for skeleton representation learning, called SkeAttnCLR, which integrates local similarity and global features for skeleton-based action representations. To achieve this, a multi-head attention mask module is employed to learn the soft attention mask features from the skeletons, suppressing non-salient local features while accentuating local salient features, thereby bringing similar local features closer in the feature space. Additionally, ample contrastive pairs are generated by expanding contrastive pairs based on salient and non-salient features with global features, which guide the network to learn the semantic representations of the entire skeleton. Therefore, with the attention mask mechanism, SkeAttnCLR learns local features under different data augmentation views. The experiment results demonstrate that the inclusion of local feature similarity significantly enhances skeleton-based action representation. Our proposed SkeAttnCLR outperforms state-of-the-art methods on NTURGB+D, NTU120-RGB+D, and PKU-MMD datasets., Comment: 7 pages, 4 figures, accepted by IJCAI 2023

It is challenging to remove rain-steaks from a single rainy image because the rain steaks are spatially varying in the rainy image. This problem is studied in this paper by combining conventional image processing techniques and deep learning based techniques. An improved weighted guided image filter (iWGIF) is proposed to extract high frequency information from a rainy image. The high frequency information mainly includes rain steaks and noise, and it can guide the rain steaks aware deep convolutional neural network (RSADCNN) to pay more attention to rain steaks. The efficiency and explain-ability of RSADNN are improved. Experiments show that the proposed algorithm significantly outperforms state-of-the-art methods on both synthetic and real-world images in terms of both qualitative and quantitative measures. It is useful for autonomous navigation in raining conditions.

The phase discontinuity caused by the geometric mutation has always been a challenging problem in phase unwrapping. To solve this problem, a fast discontinuous phase unwrapping algorithm based on orientation and transformation was proposed. In the proposed algorithm, the structure tensor of the image was used to estimate the orientation of the wrapped phase diagram. The reliable weight coefficient diagram obtained from transformation and difference calculation of the orientation diagram was taken as the weight of the weighted least square method, and used the preprocessed conjugate gradient method to solve iteratively. This algorithm could quickly find the discontinuous regions, and performed a separate phase unwrapping in the discontinuous regions, which had good recognition and unwrapping effect. The principle and implementation steps of the algorithm were described in detail, and the algorithm was verified by simulation and experimental data. The experimental results show that the root mean square error of phase unwrapping is 0.36, which proves that this algorithm can quickly and accurately unwrap the discontinuous wrapped phase.

Fully supervised skeleton-based action recognition has achieved great progress with the blooming of deep learning techniques. However, these methods require sufficient labeled data which is not easy to obtain. In contrast, self-supervised skeleton-based action recognition has attracted more attention. With utilizing the unlabeled data, more generalizable features can be learned to alleviate the overfitting problem and reduce the demand of massive labeled training data. Inspired by the MAE, we propose a spatial-temporal masked autoencoder framework for self-supervised 3D skeleton-based action recognition (SkeletonMAE). Following MAE's masking and reconstruction pipeline, we utilize a skeleton-based encoder-decoder transformer architecture to reconstruct the masked skeleton sequences. A novel masking strategy, named Spatial-Temporal Masking, is introduced in terms of both joint-level and frame-level for the skeleton sequence. This pre-training strategy makes the encoder output generalizable skeleton features with spatial and temporal dependencies. Given the unmasked skeleton sequence, the encoder is fine-tuned for the action recognition task. Extensive experiments show that our SkeletonMAE achieves remarkable performance and outperforms the state-of-the-art methods on both NTU RGB+D and NTU RGB+D 120 datasets., Comment: ICME 2023 Workshop

Existing shape from focus (SFF) techniques cannot preserve depth edges and fine structural details from a sequence of multi-focus images. Moreover, noise in the sequence of multi-focus images affects the accuracy of the depth map. In this paper, a novel depth enhancement algorithm for the SFF based on an adaptive weighted guided image filtering (AWGIF) is proposed to address the above issues. The AWGIF is applied to decompose an initial depth map which is estimated by the traditional SFF into a base layer and a detail layer. In order to preserve the edges accurately in the refined depth map, the guidance image is constructed from the multi-focus image sequence, and the coefficient of the AWGIF is utilized to suppress the noise while enhancing the fine depth details. Experiments on real and synthetic objects demonstrate the superiority of the proposed algorithm in terms of anti-noise, and the ability to preserve depth edges and fine structural details compared to existing methods.

A large number of high embankments have been constructed in highway engineering with the fast development of highway. The measurement and treatment of traffic loading-induced settlement is a research focus in recent years. This paper focuses on the measurement of the layered settlement, induced by the traffic loading, in the embankment of highway. A novel method to measure the traffic loading-induced settlement at different depths in the embankment is proposed. The practical measurement equipment, including a multipoint extensometer, rigid base, settlement detecting plate, protector tube and connecting rod, is developed. The field tests were carried out in the filling and surcharge period to verify the stability and accuracy of measurement equipment. The test results indicate that the developed equipment and the presented installation method can be capable of measuring traffic-induced settlement at different depths in the embankment of the operating highway.

Model-based single image dehazing algorithms restore images with sharp edges and rich details at the expense of low PSNR values. Data-driven ones restore images with high PSNR values but with low contrast, and even some remaining haze. In this paper, a novel single image dehazing algorithm is introduced by fusing model-based and data-driven approaches. Both transmission map and atmospheric light are initialized by the model-based methods, and refined by deep learning approaches which form a neural augmentation. Haze-free images are restored by using the transmission map and atmospheric light. Experimental results indicate that the proposed algorithm can remove haze well from real-world and synthetic hazy images.

There are shadow and highlight regions in a low dynamic range (LDR) image which is captured from a high dynamic range (HDR) scene. It is an ill-posed problem to restore the saturated regions of the LDR image. In this paper, the saturated regions of the LDR image are restored by fusing model-based and data-driven approaches. With such a neural augmentation, two synthetic LDR images are first generated from the underlying LDR image via the model-based approach. One is brighter than the input image to restore the shadow regions and the other is darker than the input image to restore the high-light regions. Both synthetic images are then refined via a novel exposedness aware saturation restoration network (EASRN). Finally, the two synthetic images and the input image are combined together via an HDR synthesis algorithm or a multi-scale exposure fusion algorithm. The proposed algorithm can be embedded in any smart phones or digital cameras to produce an information-enriched LDR image., Comment: arXiv admin note: text overlap with arXiv:2007.02042

In the highway widening project, it is of great importance to take measures to control the differential settlement between the new and old embankment, therefore improving the stability of the new embankment. This article presents a series of model tests to analyze the failure law of the new embankment and compare the reinforcement effects of different reinforcement technologies. The test results show that the new embankment is unstable because of the slope slippage under ultimate load. In the reinforcement test of gravel pile, reaming phenomenon occurred at the top of the pile, and both of the values in pavement settlement and embankment lateral deformation are similar to those of the unreinforced embankment. In the geogrid reinforcement test, the embankment settlement can be reduced by restricting the lateral deformation of the new embankment with geogrid, where the settlement is 41 % less than that of the unreinforced scenario. In the cement fly ash gravel (CFG) reinforcement test, the pavement load is transferred to a deeper layer of the embankment by CFG piles, a soil arch effect with the surrounding soil is formed to restrict the lateral deformation and slippage, and the pavement settlement is 46 % less than that of the unreinforced scenario. According to the test results, the order of effectiveness for the reinforcement methods is CFG piles > geogrid > gravel pile.

High dynamic range (HDR) image, when displayed on ordinary dynamic range devices, its visual effect will be weakened. To relieve this problem, this paper proposes a new method which combines global and local tone mapping together, to produce HDR image of better visual effect The first step is to convert a HDR image to ordinary dynamic range using either global mapping or local mapping, depending on the mean intensity of the image. Then, Curvelet transform is applied to the mapped result to obtain a set of transform coefficients which are fused. Finally, taking advantages of both global and local mapping methods, a high-quality image is obtained. Experimental results show that the processed method provides images with more details and better visual effect, compared with other mapping methods such as Reinhard tonal mapping.

The aim of the 3rd generation networks is to combine the multimedia services over Internet and provide them on wireless mobile platforms. The stability, interoperability and robustness offered by MPEG-4 (moving picture experts group) make it an ideal solution for mobile networks, and it has been selected by 3GPP (the 3rd generation partnership) and 3GPP2 for 3G phones. Burst and random bit errors introduced by fading channels degrades MPEG-4 streams. Conventional error correcting techniques cannot correct these errors. In this paper, a JSCC (joint source-channel coding) method based on MPEG-4 system and turbo coder has been proposed. The critical important MPEG-4 streams such as the BIFS (binary format for scenes), OD (object descriptors) and IOD (initial object descriptors) streams are transmitted by using TCP (transmission control protocol), and MPEG-4 audio/video ES (elementary streams) are transmitted by using UDP (user datagram protocol) over 3G channels. Transmission UDP packets is adopted UEP (unequal error protection) turbo codes according to MPEG-4 audio/video data partitioning. Simulation results show that this method can effectively improve the performance.

In this paper, a novel extreme learning machine based online multi-label classifier for real-time data streams is proposed. Multi-label classification is one of the actively researched machine learning paradigm that has gained much attention in the recent years due to its rapidly increasing real world applications. In contrast to traditional binary and multi-class classification, multi-label classification involves association of each of the input samples with a set of target labels simultaneously. There are no real-time online neural network based multi-label classifier available in the literature. In this paper, we exploit the inherent nature of high speed exhibited by the extreme learning machines to develop a novel online real-time classifier for multi-label data streams. The developed classifier is experimented with datasets from different application domains for consistency, performance and speed. The experimental studies show that the proposed method outperforms the existing state-of-the-art techniques in terms of speed and accuracy and can classify multi-label data streams in real-time., 8 pages, 7 tables, 3 figures. arXiv admin note: text overlap with arXiv:1609.00086

In this paper, a high-speed online neural network classifier based on extreme learning machines for multi-label classification is proposed. In multi-label classification, each of the input data sample belongs to one or more than one of the target labels. The traditional binary and multi-class classification where each sample belongs to only one target class forms the subset of multi-label classification. Multi-label classification problems are far more complex than binary and multi-class classification problems, as both the number of target labels and each of the target labels corresponding to each of the input samples are to be identified. The proposed work exploits the high-speed nature of the extreme learning machines to achieve real-time multi-label classification of streaming data. A new threshold-based online sequential learning algorithm is proposed for high speed and streaming data classification of multi-label problems. The proposed method is experimented with six different datasets from different application domains such as multimedia, text, and biology. The hamming loss, accuracy, training time and testing time of the proposed technique is compared with nine different state-of-the-art methods. Experimental studies shows that the proposed technique outperforms the existing multi-label classifiers in terms of performance and speed., Comment: 18 pages, 7 tables, 3 figures. arXiv admin note: text overlap with arXiv:1608.08898