SOTAVerified

Video Retrieval

The objective of video retrieval is as follows: given a text query and a pool of candidate videos, select the video which corresponds to the text query. Typically, the videos are returned as a ranked list of candidates and scored via document retrieval metrics.

Papers

Showing 110 of 486 papers

TitleStatusHype
MAGMaR Shared Task System Description: Video Retrieval with OmniEmbed0
Q2E: Query-to-Event Decomposition for Zero-Shot Multilingual Text-to-Video Retrieval0
From Play to Replay: Composed Video Retrieval for Temporally Fine-Grained VideosCode0
Leveraging Auxiliary Information in Text-to-Video Retrieval: A Review0
Learning World Models for Interactive Video Generation0
A Challenge to Build Neuro-Symbolic Video AgentsCode0
LoVR: A Benchmark for Long Video Retrieval in Multimodal ContextsCode1
Contrastive Alignment with Semantic Gap-Aware Corrections in Text-Video RetrievalCode0
Video-GPT via Next Clip DiffusionCode1
CMAWRNet: Multiple Adverse Weather Removal via a Unified Quaternion Neural Architecture0
Show:102550
← PrevPage 1 of 49Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1UMT-L (ViT-L/16)text-to-video R@190.8Unverified
2vid-TLDR (UMT-L)text-to-video R@190.2Unverified
3HiTeAtext-to-video R@185.6Unverified
4VindLUtext-to-video R@183.3Unverified
5Singularity-temporaltext-to-video R@177.6Unverified