Proposes TCSF, a compressed-domain TSG framework using I-frames, motion vectors, residuals, achieving superior performance with lower complexity.

Xiang Fang, Daizong Liu, Pan Zhou et al.

MobileVOS combines contrastive learning and knowledge distillation to enable real-time video object segmentation on mobile devices, with 32x fewer parameters and 5x faster inference.

Roy Miles, Mehmet Kerim Yucel, Bruno Manganelli et al.

SR-init assesses layer redundancy by measuring accuracy drop after stochastic re-initialization, enabling interpretable pruning with significant parameter reduction.

Hui Tang, Yao Lu, Qi Xuan

Using Flamingo as a success detector framed as VQA, achieving cross-domain generalization with minimal human annotations.

Yuqing Du, Ksenia Konyushkova, Misha Denil et al.

Retinexformer integrates one-stage Retinex framework with Transformer, achieving significant PSNR gains over 4dB on 13 datasets.

Yuanhao Cai, Hao Bian, Jing Lin et al.

VLOSS enhances open-world segmentation using omni-supervised data, surpassing MaskCLIP by 2% on LVIS v1.

Bowen Dong, Jiaxi Gu, Jianhua Han et al.

GigaGAN introduces a new architecture for text-to-image synthesis, generating 512px images in just 0.13 seconds.

Minguk Kang, Jun-Yan Zhu, Richard Zhang et al.

Visual ChatGPT integrates visual foundation models for image and language interaction.

Chenfei Wu, Shengming Yin, Weizhen Qi et al.

Utilizes pre-trained human motion diffusion models for long, multi-person, and controllable motion synthesis via sequence, parallel, and model blending strategies.

Yonatan Shafir, Guy Tevet, Roy Kapon et al.

SAN leverages a lightweight side network attached to frozen CLIP, achieving open-vocabulary semantic segmentation with only 8.4M parameters and 19× faster inference.

Mengde Xu, Zheng Zhang, Fangyun Wei et al.

DeepSORVF employs deep learning for asynchronous trajectory matching, achieving 95% accuracy in vessel data fusion under complex conditions.

Yu Guo, Ryan Wen Liu, Jingxiang Qu et al.

Survey on large-scale multi-modal pre-trained models, analyzing data, architecture, and experimental results.

Xiao Wang, Guangyao Chen, Guangwu Qian et al.

T2I-Adapter enhances text-to-image diffusion models' controllability through lightweight adapters, improving generation quality.

Chong Mou, Xintao Wang, Liangbin Xie et al.

MultiDiffusion employs multi-path optimization to enable controllable image generation without retraining, supporting diverse constraints like aspect ratio and spatial masks.

Omer Bar-Tal, Lior Yariv, Yaron Lipman et al.

PolyFormer introduces sequence-to-sequence polygon generation for referring image segmentation, achieving 5.40% and 4.52% improvements on RefCOCO+ and RefCOCOg.

Jiang Liu, Hui Ding, Zhaowei Cai et al.

ConceptFusion leverages foundation models for open-set, multimodal 3D mapping, enabling zero-shot queries with over 40% improvement in 3D IoU.

Krishna Murthy Jatavallabhula, Alihusein Kuwajerwala, Qiao Gu et al.

Proposes ViT-22B with parallel layers, QK normalization, and bias removal, achieving stable training at 22B parameters, surpassing prior models.

Mostafa Dehghani, Josip Djolonga, Basil Mustafa et al.

pix2pix-zero enables image-to-image translation without manual prompts, preserving original image structure.

Gaurav Parmar, Krishna Kumar Singh, Richard Zhang et al.

ADAPT, a Transformer-based multi-task model, jointly predicts driving actions and generates natural language explanations, achieving CIDEr 34.6 and reasoning 11.4 on BDD-X.

Bu Jin, Xinyu Liu, Yupeng Zheng et al.

Proposes Attend-and-Excite, an attention-based method to improve semantic fidelity in text-to-image diffusion models.

Hila Chefer, Yuval Alaluf, Yael Vinker et al.