Zobrazeno 1 - 10
of 204
pro vyhledávání: '"Zou, Junni"'
Autor:
Li, Han, Li, Shaohui, Ding, Shuangrui, Dai, Wenrui, Cao, Maida, Li, Chenglin, Zou, Junni, Xiong, Hongkai
Image compression for machine and human vision (ICMH) has gained increasing attention in recent years. Existing ICMH methods are limited by high training and storage overheads due to heavy design of task-specific networks. To address this issue, in t
Externí odkaz:
http://arxiv.org/abs/2407.09853
Autor:
Peng, Xinyu, Zheng, Ziyang, Dai, Wenrui, Xiao, Nuoqian, Li, Chenglin, Zou, Junni, Xiong, Hongkai
Recent diffusion models provide a promising zero-shot solution to noisy linear inverse problems without retraining for specific inverse problems. In this paper, we reveal that recent methods can be uniformly interpreted as employing a Gaussian approx
Externí odkaz:
http://arxiv.org/abs/2402.02149
Autor:
Shi, Bowen, Zhao, Peisen, Wang, Zichen, Zhang, Yuhang, Wang, Yaoming, Li, Jin, Dai, Wenrui, Zou, Junni, Xiong, Hongkai, Tian, Qi, Zhang, Xiaopeng
Vision-language foundation models, represented by Contrastive Language-Image Pre-training (CLIP), have gained increasing attention for jointly understanding both vision and textual tasks. However, existing approaches primarily focus on training model
Externí odkaz:
http://arxiv.org/abs/2401.06397
Single-cell RNA sequencing (scRNA-seq) technology provides high-throughput gene expression data to study the cellular heterogeneity and dynamics of complex organisms. Graph neural networks (GNNs) have been widely used for automatic cell type classifi
Externí odkaz:
http://arxiv.org/abs/2312.10310
Learned image compression (LIC) has gained traction as an effective solution for image storage and transmission in recent years. However, existing LIC methods are redundant in latent representation due to limitations in capturing anisotropic frequenc
Externí odkaz:
http://arxiv.org/abs/2310.16387
Autor:
Shi, Bowen, Zhang, Xiaopeng, Wang, Yaoming, Li, Jin, Dai, Wenrui, Zou, Junni, Xiong, Hongkai, Tian, Qi
Representation learning has been evolving from traditional supervised training to Contrastive Learning (CL) and Masked Image Modeling (MIM). Previous works have demonstrated their pros and cons in specific scenarios, i.e., CL and supervised pre-train
Externí odkaz:
http://arxiv.org/abs/2306.15876
JPEG images can be further compressed to enhance the storage and transmission of large-scale image datasets. Existing learned lossless compressors for RGB images cannot be well transferred to JPEG images due to the distinguishing distribution of DCT
Externí odkaz:
http://arxiv.org/abs/2303.02666
Autor:
Li, Han, Shi, Bowen, Dai, Wenrui, Zheng, Hongwei, Wang, Botao, Sun, Yu, Guo, Min, Li, Chenlin, Zou, Junni, Xiong, Hongkai
There has been a recent surge of interest in introducing transformers to 3D human pose estimation (HPE) due to their powerful capabilities in modeling long-term dependencies. However, existing transformer-based methods treat body joints as equally im
Externí odkaz:
http://arxiv.org/abs/2302.07408
Gradient coding schemes effectively mitigate full stragglers in distributed learning by introducing identical redundancy in coded local partial derivatives corresponding to all model parameters. However, they are no longer effective for partial strag
Externí odkaz:
http://arxiv.org/abs/2206.02450
Autor:
Xu, Mingxing, Li, Chenglin, Dai, Wenrui, Chen, Siheng, Zou, Junni, Frossard, Pascal, Xiong, Hongkai
Pooling and unpooling are two essential operations in constructing hierarchical spherical convolutional neural networks (HS-CNNs) for comprehensive feature learning in the spherical domain. Most existing models employ downsampling-based pooling, whic
Externí odkaz:
http://arxiv.org/abs/2205.15571