Video portrait segmentation(VPS), aiming at segmenting prominent foreground portraits from video frames, has received much attention in recent years. However, the simplicity of existing VPS datasets leads to a limitat...
详细信息
Video portrait segmentation(VPS), aiming at segmenting prominent foreground portraits from video frames, has received much attention in recent years. However, the simplicity of existing VPS datasets leads to a limitation on extensive research of the task. In this work, we propose a new intricate large-scale multi-scene video portrait segmentation dataset MVPS consisting of 101 video clips in 7 scenario categories,in which 10843 sampled frames are finely annotated at the pixel level. The dataset has diverse scenes and complicated background environments, which is the most complex dataset in VPS to our best *** the observation of a large number of videos with portraits during dataset construction, we find that due to the joint structure of the human body, the motion of portraits is part-associated, which leads to the different parts being relatively independent in motion. That is, the motion of different parts of the portraits is imbalanced. Towards this imbalance, an intuitive and reasonable idea is that different motion states in portraits can be better exploited by decoupling the portraits into parts. To achieve this, we propose a part-decoupling network(PDNet) for VPS. Specifically, an inter-frame part-discriminated attention(IPDA)module is proposed which unsupervisedly segments portrait into parts and utilizes different attentiveness on discriminative features specified to each different part. In this way, appropriate attention can be imposed on portrait parts with imbalanced motion to extract part-discriminated correlations, so that the portraits can be segmented more accurately. Experimental results demonstrate that our method achieves leading performance with the comparison to state-of-the-art methods.
A scheme to distinguish the interfered image frequencies between electro-optic modulated dual combs is proposed. Measurement of the absorptive spectrum of acetylene (C2H2) based on a simple setup is demonstrated. ...
详细信息
Two-color pump-probe measurement of silicon waveguides is realized using an all-PM dual-comb fiber laser. High temporal resolution results simultaneously reveal the interplay of multiple nonlinear dynamics of ps-to-ns...
详细信息
Broadband frequency responses of silicon photonic photodetectors are experimentally investigated under linear and saturated conditions. It is shown that long photodetectors could have better high frequency outputs bey...
详细信息
An compact inline power monitor with a short Ge segment over a silicon multimode interference (MMI) structure is proposed and fabricated. With ∼0.36dB loss, 15mA/W responsivity at 1550nm, it is useful for high-densit...
Large 2219 Al-Cu alloy aerospace integral components suffer from long-term stress relaxation aging(SRA)due to complex temperature and stress loads during aging treatment/forming and service process,which makes it diff...
详细信息
Large 2219 Al-Cu alloy aerospace integral components suffer from long-term stress relaxation aging(SRA)due to complex temperature and stress loads during aging treatment/forming and service process,which makes it difficult to ensure their appropriate residual stress and excellent mechanical and service ***,the research is limited to a thorough understanding of macroscopic and microscopic features and underlying mechanisms of the long-term SRA under multivariable aging ***-fore,this study investigated macroscopic and microscopic features of long-term SRA under different tem-peratures(120 ℃ to 190 ℃),initial stress levels(100 MPa to 250 MPa)and durations(0 h to 50 h)through stress relaxation curves,metallographic traits,Vickers hardness,tensile performance,disloca-tions and phases of *** the basis of experimental outcomes,the comprehensive mecha-nisms beneath SRA were unraveled through dislocation theory,multiphase strengthening mechanisms and thermodynamics,where the interplays of stress relaxation behavior with age-hardening response were taken into *** results showed elevations in the rate of stress reduction as the tem-perature and initial stress *** an initial stress greater than the yield stress of alloy,a marked in-crease in stress relaxation was found,and the mechanisms transform from the intragranular motion of dislocations and diffusion of grain boundaries to the intragranular and intergranular motion of disloca-tions and migration of grain *** stress reduction rate rose sharply when the temperature exceeded 175 ℃,and the dislocation movement mechanisms transform from gliding to climbing of *** relaxation is in nature progressive transformation of strain from elastic into a permanently inelastic state via the motion of dislocations,leading to the decrease of movable dislocations and the increase of immovable dislocations with more stable *** age hardening i
Glass wool defect detection is a key part of product quality assessment in the glass wool production process, yet few studies have been reported in this area. We propose a glass wool defect dataset named GWD, and also...
详细信息
Light field cameras hold significant value in applications such as depth estimation, 3D video acquisition, and image super-resolution. Compared to traditional single-image super-resolution methods, light field images ...
详细信息
Though feature-alignment based Domain Adaptive Object Detection (DAOD) methods have achieved remarkable progress, they ignore the source bias issue, i.e., the detector tends to acquire more source-specific knowledge, ...
详细信息
Though feature-alignment based Domain Adaptive Object Detection (DAOD) methods have achieved remarkable progress, they ignore the source bias issue, i.e., the detector tends to acquire more source-specific knowledge, impeding its generalization capabilities in the target domain. Furthermore, these methods face a more formidable challenge in achieving consistent classification and localization in the target domain compared to the source domain. To overcome these challenges, we propose a novel Distillation-based Source Debiasing (DSD) framework for DAOD, which can distill domain-agnostic knowledge from a pre-trained teacher model, improving the detector's performance on both domains. In addition, we design a Target-Relevant Object Localization Network (TROLN), which can mine target-related localization information from source and target-style mixed data. Accordingly, we present a Domain-aware Consistency Enhancing (DCE) strategy, in which these information are formulated into a new localization representation to further refine classification scores in the testing stage, achieving a harmonization between classification and localization. Extensive experiments have been conducted to manifest the effectiveness of this method, which consistently improves the strong baseline by large margins, outperforming existing alignment-based works. Copyright 2024 by the author(s)
Eagle,a representative species in the raptor world,has the sharpest visual acuity among all *** reputation of the“clairvoyance”is employed to describe an *** excellent visual skills of eagles depend on their unique ...
详细信息
Eagle,a representative species in the raptor world,has the sharpest visual acuity among all *** reputation of the“clairvoyance”is employed to describe an *** excellent visual skills of eagles depend on their unique eye structures and special visual *** powerful vision perception mechanisms of the eagle bring abundant inspiration for traditional visual *** eagle eye vision technology provides a creative way to solve visual perception issues of“Knowing What is Where by Seeing.”The theoretical research and practical works of eagle vision would contribute to the development of machine vision,or even artificial intelligence(AI)in the real ***,eagle eye vision also provides feasible ideas for the popularization of new concepts in the virtual world in the future.
暂无评论