Significant advancements have beenwitnessed in visual tracking applications leveragingViT in recent years,mainly due to the formidablemodeling capabilities of Vision Transformer(ViT).However,the strong performance of ...Significant advancements have beenwitnessed in visual tracking applications leveragingViT in recent years,mainly due to the formidablemodeling capabilities of Vision Transformer(ViT).However,the strong performance of such trackers heavily relies on ViT models pretrained for long periods,limitingmore flexible model designs for tracking tasks.To address this issue,we propose an efficient unsupervised ViT pretraining method for the tracking task based on masked autoencoders,called TrackMAE.During pretraining,we employ two shared-parameter ViTs,serving as the appearance encoder and motion encoder,respectively.The appearance encoder encodes randomly masked image data,while the motion encoder encodes randomly masked pairs of video frames.Subsequently,an appearance decoder and a motion decoder separately reconstruct the original image data and video frame data at the pixel level.In this way,ViT learns to understand both the appearance of images and the motion between video frames simultaneously.Experimental results demonstrate that ViT-Base and ViT-Large models,pretrained with TrackMAE and combined with a simple tracking head,achieve state-of-the-art(SOTA)performance without additional design.Moreover,compared to the currently popular MAE pretraining methods,TrackMAE consumes only 1/5 of the training time,which will facilitate the customization of diverse models for tracking.For instance,we additionally customize a lightweight ViT-XS,which achieves SOTA efficient tracking performance.展开更多
Chemokines belong to a large family of inflammatory cytokines responsible for migration and accumulation of leukocytes at inflammatory sites.Over the past decade,accumulating evidence indicated a crucial role for chem...Chemokines belong to a large family of inflammatory cytokines responsible for migration and accumulation of leukocytes at inflammatory sites.Over the past decade,accumulating evidence indicated a crucial role for chemokines and chemokine receptors in the pathophysiology of rheumatoid arthritis(RA).RA is a chronic autoimmune disease in which the synovial tissue is heavily infiltrated by leukocytes.Chemokines play an important role in the infiltration,localization,retention of infiltrating leukocytes and generation of ectopic germinal centers in the inflamed synovium.Recent evidence also suggests that identification of inhibitors directly targeting chemokines or their receptors may provide a novel therapeutic strategy in RA.Traditional Chinese medicines(TCMs) have a long history in the treatment of inflammatory joint disease.The basis for the clinical benefits of TCM remains largely unclear.Our studies have led to the identification of numerous novel chemokine/chemokine receptor inhibitors present in anti-inflammatory TCMs.All of these inhibitors were previously reported by other researchers to have anti-arthritic effect,which may be attributable,at least in part,to their inhibitory effect on chemokine and/or chemokine receptor.Therefore,identification of agents capable of targeting chemokine/chemokine receptor interactions has suggested a mechanism of action for several TCM components and provided a means of identifying additional anti-RA TCM.Thus,this approach may lead to the discovery of new inhibitors of chemokines or chemokine receptors that can be used to treat diseases associated with inappropriately overactive chemokine mediated inflammatory reactions.Cellular & Molecular Immunology.2004;1(5):336-342.展开更多
基金supported in part by National Natural Science Foundation of China(No.62176041)in part by Excellent Science and Technique Talent Foundation of Dalian(No.2022RY21).
文摘Significant advancements have beenwitnessed in visual tracking applications leveragingViT in recent years,mainly due to the formidablemodeling capabilities of Vision Transformer(ViT).However,the strong performance of such trackers heavily relies on ViT models pretrained for long periods,limitingmore flexible model designs for tracking tasks.To address this issue,we propose an efficient unsupervised ViT pretraining method for the tracking task based on masked autoencoders,called TrackMAE.During pretraining,we employ two shared-parameter ViTs,serving as the appearance encoder and motion encoder,respectively.The appearance encoder encodes randomly masked image data,while the motion encoder encodes randomly masked pairs of video frames.Subsequently,an appearance decoder and a motion decoder separately reconstruct the original image data and video frame data at the pixel level.In this way,ViT learns to understand both the appearance of images and the motion between video frames simultaneously.Experimental results demonstrate that ViT-Base and ViT-Large models,pretrained with TrackMAE and combined with a simple tracking head,achieve state-of-the-art(SOTA)performance without additional design.Moreover,compared to the currently popular MAE pretraining methods,TrackMAE consumes only 1/5 of the training time,which will facilitate the customization of diverse models for tracking.For instance,we additionally customize a lightweight ViT-XS,which achieves SOTA efficient tracking performance.
文摘Chemokines belong to a large family of inflammatory cytokines responsible for migration and accumulation of leukocytes at inflammatory sites.Over the past decade,accumulating evidence indicated a crucial role for chemokines and chemokine receptors in the pathophysiology of rheumatoid arthritis(RA).RA is a chronic autoimmune disease in which the synovial tissue is heavily infiltrated by leukocytes.Chemokines play an important role in the infiltration,localization,retention of infiltrating leukocytes and generation of ectopic germinal centers in the inflamed synovium.Recent evidence also suggests that identification of inhibitors directly targeting chemokines or their receptors may provide a novel therapeutic strategy in RA.Traditional Chinese medicines(TCMs) have a long history in the treatment of inflammatory joint disease.The basis for the clinical benefits of TCM remains largely unclear.Our studies have led to the identification of numerous novel chemokine/chemokine receptor inhibitors present in anti-inflammatory TCMs.All of these inhibitors were previously reported by other researchers to have anti-arthritic effect,which may be attributable,at least in part,to their inhibitory effect on chemokine and/or chemokine receptor.Therefore,identification of agents capable of targeting chemokine/chemokine receptor interactions has suggested a mechanism of action for several TCM components and provided a means of identifying additional anti-RA TCM.Thus,this approach may lead to the discovery of new inhibitors of chemokines or chemokine receptors that can be used to treat diseases associated with inappropriately overactive chemokine mediated inflammatory reactions.Cellular & Molecular Immunology.2004;1(5):336-342.
