ObjectFolder 2.0 uses implicit neural representations to build a large-scale multisensory dataset, enabling effective Sim2Real transfer for robotics and vision tasks.

Ruohan Gao, Zilin Si, Yen-Yu Chang et al.

Proposed an autoregressive 3D shape generation method using Transformers, achieving efficient generation via semantically aligned shape composition sequences.

An-Chieh Cheng, Xueting Li, Sifei Liu et al.

GraftNet leverages large-scale pretrained features with shallow adaptation to improve cross-domain stereo matching, reducing error rates from 19.5% to 5.34%.

Biyang Liu, Huimin Yu, Guodong Qi

OakInk combines object affordance and human interaction data, with 50,000 instances, advancing pose estimation and interaction generation.

Lixin Yang, Kailin Li, Xinyu Zhan et al.

OpenDet method enhances open-set object detection by expanding low-density latent regions, reducing errors by 25%-35%.

Jiaming Han, Yuqiang Ren, Jian Ding et al.

Introduced a spatio-temporal grounded audio-visual network, excelling on the MUSIC-AVQA dataset.

Guangyao Li, Yake Wei, Yapeng Tian et al.

UMT framework significantly improves video moment retrieval and highlight detection, excelling on the QVHighlights dataset.

Ye Liu, Siyuan Li, Yang Wu et al.

VideoMAE employs extremely high masking ratios (90%-95%) for self-supervised pretraining, significantly boosting video recognition accuracy.

Zhan Tong, Yibing Song, Jue Wang et al.

This paper introduces Visual Prompt Tuning (VPT), which inserts less than 1% trainable prompts into the input space of frozen vision transformers, outperforming full fine-tuning on 20 out of 24 tasks with significantly fewer parameters.

Menglin Jia, Luming Tang, Bor-Chun Chen et al.

Survey of Vision-and-Language Navigation tasks, methods, and future directions, emphasizing datasets and evaluation metrics.

Jing Gu, Eliana Stefani, Qi Wu et al.

CREStereo employs a hierarchical recurrent network with adaptive correlation, achieving top performance on Middlebury and ETH3D benchmarks with significant detail preservation.

Jiankun Li, Peisen Wang, Pengfei Xiong et al.

ELIC introduces uneven space-channel context modeling, leveraging energy compaction for fast, high-quality image compression with rapid preview decoding.

Dailan He, Ziming Yang, Weikun Peng et al.

This paper introduces a single-image facial detail recovery method using implicit surface representation and StyleGAN-based texture completion, achieving state-of-the-art results.

Xingyu Ren, Alexandros Lattas, Baris Gecer et al.

TensoRF employs tensor decomposition of 4D feature tensors for fast, compact scene radiance field reconstruction, outperforming NeRF in speed and size.

Anpei Chen, Zexiang Xu, Andreas Geiger et al.

AutoSDF uses autoregressive priors for 3D shape completion, reconstruction, and generation, surpassing existing methods.

Paritosh Mittal, Yen-Chi Cheng, Maneesh Singh et al.

Self-supervised MAE pretraining on large-scale real images improves pixel-based robot control, outperforming supervised encoders by up to 80%.

Tete Xiao, Ilija Radosavovic, Trevor Darrell et al.

This study reveals FID's high dependence on ImageNet classes, its manipulability, and proposes Top-N classification histogram matching to analyze biases.

Tuomas Kynkäänniemi, Tero Karras, Miika Aittala et al.

DINO integrates contrastive denoising, mixed query initialization, and look-forward twice, achieving 49.4AP in 12 epochs, outperforming DN-DETR.

Hao Zhang, Feng Li, Shilong Liu et al.

Proposes SSCD, a self-supervised contrastive descriptor for image copy detection, outperforming baselines by 48% on DISC2021.

Ed Pizzi, Sreya Dutta Roy, Sugosh Nagavara Ravindra et al.

Proposes a continuous conditional GAN (CC-GAN) for 3D point cloud generation, achieving explicit size control with low regression error (0.28%) and superior quality (FPD 1.5290).

Larissa T. Triess, Andre Bühler, David Peter et al.