The recent advances of hardware technology have made the intelligent analysis equipped at front-end with deep learning more prevailing and practical. To better enable sensing front-end, instead compressing transmitting visual signals or ultimately utilized top-layer features, we propose to compactly represent convey intermediate-layer features high generalization capability, facilitate collaborating approach between front cloud ends. This strategy enables a good balance among computational...

With the unprecedented success of deep learning in computer vision tasks, many cloud-based visual analysis applications are powered by models. However, models also characterized with high computational complexity and task-specific, which may hinder large-scale implementation conventional data communication paradigms. To enable a better balance among bandwidth usage, load generalization capability for cloud-end servers, we propose to compress transmit intermediate features instead signals...

3D sensing and content capturing have made significant progress in recent years the MPEG standardization organization is launching a new project on immersive media with point cloud compression (PCC) as one key corner stone. In this work, we introduce binary tree based partition explore graph signal processing tools, especially transform optimized Laplacian sparsity, to achieve better energy compaction efficiency. The resulting rate-distortion operating points are convex-hull over existing...

Laser welding, as an important material processing technology, has been widely used in various fields of industry. In most industrial welding production and processing, high precision is required for parameters fixed work pieces. However, the process laser serious heat transfer effect will bring unpredictable deviations, even a small deviation lead to defects, which affect quality welded products. Traditional nondestructive testing methods have used, but they proved some limitations....

Most of the existing laser welding process monitoring technologies focus on detection post-engineering defects, but in mass production electronic equipment, such as metal plates, real-time identification defect has more important practical significance. The data set is often difficult to build and there not enough experimental data, which hinder applications data-driven method. In this paper, an intelligent diagnosis method based auxiliary classifier generative adversarial networks (ACGAN)...

The state-of-the-art high efficiency video coding (HEVC) standard provides a significant improvement relative to H.264/AVC, with almost 50% bitrate reduction. However, there are still requirements further improve the performance of HEVC. In this paper, we reveal some problems in both intra and inter prediction by giving detailed observations, solve them exploring spatial temporal correlations. First, it is noteworthy that large boundary distortions blocks arise from simplistic extrapolation...

The baseline profile of the third generation Audio Video coding Standard (AVS3) was finalized in March 2019, and high is expected to be issued 2021. In AVS3 profile, various key technologies on structure, prediction, transform, loop filter are newly adopted. achieves a significant improvement efficiency relative previous video standards. experimental results show that bitrate savings 23.52% 22.25% averagely compared with AVS2 HEVC, respectively. computational complexity rises inevitably due...

The second generation of the audio video coding standard (AVS2), which has been issued as IEEE 1857.4, doubles efficiency AVS1 and H.264/AVC. However, advanced techniques in AVS2 also dramatically increase computational complexity. research on a commercial encoder for is still at an early stage. In this paper, we propose first fast intra-encoding platform AVS2, term iAVS2. uses numerous speedup methods, including code optimization, single-instruction multiple-data acceleration, algorithms...

MPEG has produced standards that have provided the industry with best video compression technologies. To address diverse Internet needs, issued a Call for Proposals (CfP) coding (IVC) in July, 2011. The anticipation is any patent declaration associated baseline profile of this standard will indicate owner prepared to grant free charge license an unrestricted number applicants worldwide. Three codecs responded CfP: Web (WVC), browsers (VCB), and IVC. WVC fact AVC baseline, VCB uses same tools...

Panoramic video provides an immersive experience by presenting a 360° spherical content. Due to the limitations of coding and storage technology, panoramic needs be projected onto two-dimensional plane for encoding. In this paper, we propose polar square projection scheme. We project area near poles sphere into two planes latitude circle on is squares plane, in addition, rest rectangle means equal projection. Experimental results show our proposed can obtain gain 11.63% BD-rate compared...

Merge prediction is a practical inter-technique in HEVC, which can significantly improve the coding efficiency, especially for homogeneous regions video sequences. In this paper, motion aided merge mode (MAMM) proposed to achieve better trade-off between accuracy and bit rate. Different from traditional MAMM accomplished by small obtained searching specific search region. The range comprised of number points with high occurrence possibilities. vector difference (MVD) coded Huffman table...

After the works on High Efficiency Video Coding (HEVC) standard, standard organizations continued to study next generation of video coding named Versatile (VVC). The compression capacity VVC is expected be substantially improved relative current HEVC by evolving potential tools greatly. Transform a key technique for efficiency, and core experiment 6 (CE6) in JVET established explore transform-related tools. In this paper, we propose novel signal-independent separable transform based...

This paper presents a group of novel intra boundary filters to refine the prediction blocks. The existing only adjust blocks for few special modes. While proposed method extends filtering all modes, and allows modify up 6 rows columns at boundaries A are designed by comprehensively considering spatial similarity between reference samples predicted pixels, as well statistical characters distortion. In addition, low computational complexity is also essential make simple effective. By involving...

Graph representation learning has emerged as a powerful tool for preserving graph topology when mapping nodes to vector representations, enabling various downstream tasks such node classification and community detection. However, most current neural network models face the challenge of requiring extensive labeled data, which limits their practical applicability in real-world scenarios where data is scarce. To address this challenge, researchers have explored Contrastive Learning (GCL),...

Compared with AVS2 and HEVC, the compression performance of 3rd generation audio video coding standard (AVS3) has been greatly improved. For preliminary application this emerging standard, an efficient implementation software-based decoder is necessary before decoding chips are widely used. However, real-time UHD streams still a big challenge for AVS3 standard. In paper, we proposed fast design AVS3, consisting framework optimization, data structure module optimizations, SIMD parallel...

This paper presents a novel intra prediction method for inter pictures (i.e. P and B pictures), denominated enhanced (EIP). The traditional only uses the reconstructed pixels to left above derive blocks. While proposed combines left-above right-below strengthen efficiency of blocks in pictures. For accessing below right coding units (CUs), encoding decoding structures are adjusted guarantee that all CUs before CUs. With more available reference samples around CUs, EIP achieves better...

3D sensing and content capture have made significant progress in recent years the MPEG standardization organization is launching a new project on immersive media with point cloud compression (PCC) as one key corner stone. In this work, we introduce binary tree based partition explore graph signal processing tools, especially transform optimized Laplacian sparsity, to achieve better energy compaction efficiency. The resulting rate-distortion operating points are convex-hull over existing...

After the works on state-of-the-art High Efficiency Video Coding (HEVC) standard, standard organizations continued to study potential video coding technologies for next generation of named Versatile (VVC). Transform is a key technique compression efficiency, and core experiment 6 (CE6) carried out explore transform related tools. In this paper, we propose novel separable based Karhunen-Loève (KLT) eliminate horizontal vertical correlations in residual samples intra coding. proposed method,...

Internet Video Coding (IVC) has been developed in MPEG by combining well-known existing technology elements and new coding tools with royalty-free declarations. In June 2015, IVC project was approved as ISO/IEC 14496-33 (MPEG- 4 Coding). It is believed that this standard can be highly beneficial for video services the domain. This paper evaluates objective subjective performances of comparing it against Web (WVC), Browsers (VCB) AVC High Profile. Experimental results show IVC's compression...

In this paper, a novel self-adaptive curve, based on human visual model (HVM), is proposed for recovering details from low dynamic range (LDR) digital photographs, which are under-exposed or over-exposed both. order to improve the perceptual visibility, we utilize HVM construct our method, able take advantage of entire enhance contrast images. Extensive experiments demonstrate that method consistently achieves satisfying results unwell-exposed LDR photographs.