ActivityNet: A large-scale video benchmark for human activity understanding

By Fabian Caba Heilbron, Victor Escorcia, Bernard Ghanem, Juan Carlos Niebles. 2015

3,001 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Evaluation & Benchmarks

Citations in the corpus

SWAG: A Large-Scale Adversarial Dataset for Grounded Commonsense Inference (2018, methodology): “secutive video captions from ActivityNet Captions (Krishna et al., 2017; Heilbron et al., 2015) and the Large Scale Movie Description Challenge (LSMDC; Rohrbach et al.”
End-to-End Learning of Visual Representations From Uncurated Instructional Videos (2019, background): “Notably, the largest action datasets today are still orders of magnitude smaller (UCF101 [36], ActivityNet [7]) compared to large image datasets, they often contain label noise and target specific domains such as sports (Sports1M [20]).”
MERLOT: Multimodal Neural Script Knowledge Models (2021, methodology): “We report results on 12 video reasoning tasks: TVQA [64], TVQA(+) [65], VLEP [66], MSRVTT-QA [117], MSRVTT-Multichoice [121], LSMDC-Multichoice, LSMDC ﬁll-in-the-blank QA [110, 92], ActivityNetQA [122, 45], TGIFQA [52], and DramaQA [23].”
VaTeX: A Large-Scale, High-Quality Multilingual Dataset for Video-and-Language Research (2019, methodology): “[28] first built an action list by combining previous video datasets [24, 31, 53, 3, 62], and then searched the videos from YouTube for candidates, which eventually were filtered by Amazon Mechanical Turkers.”
Ego4D: Around the World in 3,000 Hours of Egocentric Video (2021, background): “[61] Bernard Ghanem Fabian Caba Heilbron, Victor Escorcia and Juan Carlos Niebles.”
Rescaling Egocentric Vision: Collection, Pipeline and Challenges for EPIC-KITCHENS-100 (2020, methodology): “ActivityNet [27] is the most popular video challenge, evaluated for localisation, dense captioning [31] and object detection [32].”
The Kinetics Human Action Video Dataset (2017, methodology): “Clips Total Videos HMDB-51 [15] 2011 51 min 102 6,766 3,312 UCF-101 [20] 2012 101 min 101 13,320 2,500 ActivityNet-200 [3] 1: Statistics for recent human action recognition datasets.”
Dense-Captioning Events in Videos (2017, background): “To evaluate our model and benchmark progress in dense-captioning events, we introduce the ActivityNet Cap-tions dataset 1 .”
Rethinking Spatiotemporal Feature Learning: Speed-Accuracy Trade-offs in Video Classification (2017, background): “However, the recent creation of Sports-1M [5], Kinetics [6], Something-something [7], ActivityNet [8], Charades [9], etc. has partially removed that impediment.”
A Joint Sequence Fusion Model for Video Question Answering and Retrieval (2018, background): “Video classification [14,15,16,17,18] can be a candidate solution, but tagging only a few labels to a video may be insufficient to fully relate multiple latent events in the video to a language description.”
Hollywood in Homes: Crowdsourcing Data Collection for Activity Understanding (2016, background): “In order to scale up the learning and the complexity of the data, recent approaches have instead tried collecting video datasets by downloading videos from Internet.”
End-to-End Dense Video Captioning with Masked Transformer (2018, positive): “Both networks are pre-trained on the ActivityNet dataset [5] for the action recognition task.”
YouTube-8M: A Large-Scale Video Classification Benchmark (2016, methodology): “…activities and 1 M videos, the YFCC-100M [34], with 800 K videos and raw metadata (titles, descriptions, tags) for some of them, the FCVID [17] dataset of 91 , 223 videos manually annotated with 239 categories, and ActivityNet [12], with ∼ 200 human activity classes and a few thousand videos.”
AVA: A Video Dataset of Spatio-Temporally Localized Atomic Visual Actions (2017, background): “ActivityNet [6], THUMOS [16], MultiTHUMOS [36] and Charades [32] use large numbers of untrimmed videos, each containing multiple actions, obtained either from YouTube (ActivityNet, THUMOS, MultiTHUMOS) or from crowdsourced actors (Charades).”
NTU RGB+D 120: A Large-Scale Benchmark for 3D Human Activity Understanding (2019, background): “Unlike RGB video-based activity analysis [5], [6], [7], [8], 3D action recognition suffers from the lack of large-scale benchmark datasets.”
Learning Spatio-Temporal Features with 3D Residual Networks for Action Recognition (2017, neutral): “3) [4] and Kinetics datasets [12].”
CLEVRER: CoLlision Events for Video REpresentation and Reasoning (2019, background): “With the availability of large-scale video datasets (Caba Heilbron et al., 2015; Kay et al., 2017), joint video and language understanding tasks have received much interest.”
vCLIMB: A Novel Video Class Incremental Learning Benchmark (2022, methodology): “3K videos from 101 classes, (2) Kinetics [6], a large-scale video dataset with over 300K short clips distributed over 400 action classes, and (3) ActivityNet [5], which can be used for trimmed and untrimmed activity classification and contains 20K videos from 200 activity classes.”
RGB-D-based action recognition datasets: A survey (2016, methodology): “A newly released 2D dataset [108] on action recognition contains 203 distinct action classes in 849 hours of video recording.”
Video Representation Learning Using Discriminative Pooling (2018, neutral): “They are samples from (i) ActivityNet dataset [4] unrelated to HMDB-51, (ii) UCF-101 dataset unrelated to HMDB51, (iii) Thumos Challenge background sequence(2), and (iv) synthesized random white noise image sequences.”