https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F1984321%2Fee10abf5409ea4eaaad3dfaa9514a4bb%2FScreenshot_2021-08-06_at_16.15.03.png?generation=1694441423300452&alt=media" alt="">

Video Action Recognition : Kinetics 400

The dataset contains 400 human action classes, with at least 400 video clips for each action. Each clip lasts around 10s and is taken from a different YouTube video. The actions are human focussed and cover a broad range of classes including human-object interactions such as playing instruments, as well as human-human interactions such as shaking hands. Homepage.

License

The kinetics dataset is licensed by Google Inc. under a Creative Commons Attribution 4.0 International License. Published. May 22, 2017.

The Kinetics dataset is a large-scale, high-quality dataset for human action recognition in videos. The dataset consists of around 500,000 video clips covering 600 human action classes with at least 600 video clips for each action class. Each video clip lasts around 10 seconds and is labeled with a single action class. The videos are collected from YouTube.

MD5 checksum: kinetics400.zip: 33224b5b77c634aa6717da686efce2d4 kinetics400_validation.zip: 013358d458477d7ac10cebb9e84df354

JackWong0911/kinetic-400_450samples dataset hosted on Hugging Face and contributed by the HF Datasets community

Dataset Description

Original source: https://www.deepmind.com/open-source/kinetics

  innat/KineticsTop5

A small set from Kinetics-400. It contains 5 classes. {0: 'opening_bottle', 1: 'squat', 2: 'reading_book', 3: 'sneezing', 4: 'reading_newspaper'}

kinetics_top5.zip: No internal data drop. kinetics_top5_tiny.zip: Internal data drop.

The Kinetics-400, UCF101, HMDB51, Something-Something V1, and Something-Something V2 datasets are used for evaluating the performance of the Bi-Calibration Networks.

The Kinetics-600 is a large-scale action recognition dataset which consists of around 480K videos from 600 action categories. The 480K videos are divided into 390K, 30K, 60K for training, validation and test sets, respectively. Each video in the dataset is a 10-second clip of action moment annotated from raw YouTube video. It is an extensions of the Kinetics-400 dataset.

kiyoonkim/kinetics-400-splits dataset hosted on Hugging Face and contributed by the HF Datasets community

Kinetics-400 Video Dataset

This dataset is derived from the Kinetics-400 dataset, which is released under the Creative Commons Attribution 4.0 International License (CC BY 4.0).

  Attribution

This dataset is derived from:

Original Dataset: Kinetics-400 Original Authors: Will Kay, Joao Carreira, Karen Simonyan, Brian Zhang, Chloe Hillier, Sudheendra Vijayanarasimhan, Fabio Viola, Tim Green, Trevor Back, Paul Natsev, Mustafa Suleyman, Andrew Zisserman Original Paper: "The… See the full description on the dataset page: https://huggingface.co/datasets/liuhuanjim013/kinetics400.

The Kinetics dataset is a large-scale human action dataset, which consists of 400 action classes where each category has more than 400 videos.

The dataset used in the paper is Kinetics-400 and Something-Something-V2.

Oxidation kinetic experiments with various crude oil types show two reaction peaks at about 250{degree}C (482{degree}F) and 400{degree}C (725{degree}F). These experiments lead to the conclusion that the fuel during high temperature oxidation is an oxygenated hydrocarbon. A new oxidation reaction model has been developed which includes two partially-overlapping reactions: namely, low-temperature oxidation followed by high-temperature oxidation. For the fuel oxidation reaction, the new model includes the effects of sand grain size and the atomic hydrogen-carbon (H/C) and oxygen-carbon (O/C) ratios of the fuel. Results based on the new model are in good agreement with the experimental data. Methods have been developed to calculate the atomic H/C and O/C ratios. These methods consider the oxygen in the oxygenated fuel, and enable a direct comparison of the atomic H/C ratios obtained from kinetic and combustion tube experiments. The finding that the fuel in kinetic tube experiments is an oxygenated hydrocarbon indicates that oxidation reactions are different in kinetic and combustion tube experiments. A new experimental technique or method of analysis will be required to obtain kinetic parameters for oxidation reactions encountered in combustion tube experiments and field operations.

Hydrogen (H2) activation is fundamental in catalysis. Single-atom catalysts (SACs) can be highly selective in many reactions invoking H2 activation due to their tunable geometric and electronic properties. In this work, we employ density functional theory (DFT) and microkinetic modeling (MKM) to study H2 activation (adsorption, dissociation, and diffusion) on the dehydroxylated (101) facet of anatase TiO2 (corresponding to a water-free reaction environment) over 14 single-atom transition metals from 3d to 5d (Fe, Co, Ni, Cu, Zn, Ru, Rh, Pd, Ag, Cd, Os, Ir, Pt, and Au) and Sn. The stability of intermediates from the dissociative adsorption of H2 is first evaluated, and linear scaling relationships are explored for H···H dissociation and diffusion. We find that linear scalings are generally inadequate for H2 activation. MKM simulations show that H2 activation over the SA/TiO2 sites occurs under kinetic control at moderate temperatures (below 400 K). Thermodynamically preferred H–H splitting states are achieved via kinetically favored splitting followed by subsequent diffusion steps. Overall, adsorption is faster for SA sites with weaker SA–H interactions as more empty surface sites are exposed. H–H dissociation takes place by following the path with the lowest barrier but may lead to metastable products, where the most stable surface intermediates are reached via H diffusion, potentially leading to site poisoning. Up to 400 K, the system generally cannot reach steady state within 3 h, leading to diverse hydride (M–H) or OH sites that depend on the SA, the temperature, and exposure time. Temperature-programmed desorption (TPD) simulations reveal that the observed H2 desorption peaks strongly correlate with the exposure temperature and the SA’s chemical nature, further demonstrating the importance of kinetics in H2 activation by SA sites.

Significant accomplishment this quarter have been demonstration that: the procedures that we are using in the tubing bomb experiments yield results comparable to data published by Rindt et al. from North Dakota Energy Research Center (UNDERC,5935 ); showing that the ZAP-2 recycle solvent was a very poor donor solvent; and presentation of the automation of the size exclusion chromatograph-gas chromatograph at the Fall American Chemical Society meeting. We also present data on a mixing study conducted with ZAP-2 recycle solvent and ZAP-2 lignite. Even though the yields are poor, the same yields within experimental error are obtained at 400, 600 and 700 cpm. These data substantiate data published by Gallakota and guin, that shakers speeds equal to or greater than 400 cpm are sufficient and necessary to obtain good mixing. We also present data with anthracene oil and ZAP-2 lignite in which high yields are obtained. The high yields with anthracene oil and ZAP-2 lignite as compared to the low yields with ZAP-2 recycle solvent indicate that the ZAP-2 recycle solvent supplied by UNDERC was in exceeding poor hydrogen donor. In this quarterly, data for Big Brown-2 lignite with anthracene oil (A04) that are comparable to data published by Rindt et al. are presented. Data presented in this quarterly also seem to suggest that temperature greater than 400 C and pressures of 27 to 30 MPa are needed to obtain good distillable yields. Other accomplishments for the year have been the installation and commissioning of the Finnigan Ion Trap Mass Spectrometer Detector, the development of a Reactor and Kinetic Model for the Up Flow Tubular Reactor at UNDERC by using data generated by that unit. 4 refs., 7 figs., 14 tabs.

Progress reports are presented for the following tasks: (1) batch liquefaction of ZAP-2 lignite with a UNDERC-PDU-recycle solvent; (2) estimation of parameters in a bubble column reactor for coal liquefaction; and (3) analysis of the coal derived products. During this quarter the emphasis has been on the refining of the experimental techniques for conducting the tubing bomb reactor experiments, analyzing the products by size exclusion chromatography (SEC) and size exclusion chromatography combined with gas chromatography (SEGC). The contents of the tubing bomb microreactors are mixed by insertion of a short 1.5 mm stainless steel rod inside the reactor. The reactors are agitated by shaking at 400 cpm. A cold experiment with mineral oil in a glass tube was used to test the degree of mixing. Data are reported that were obtained by agitating the reactors at 220 cpm which was believed to be adequate. However, because the data show repolymerization, the system was redesigned to operate at 400 cpm. The data which show polymerization and condensation of THF soluble material were also taken at 793/sup 0/K. This high temperature combined with the fact that the product slurry is probably not a very good hydrogen donor solvent could also have caused the polymerization of the THF solubles. The SEGC analysis procedure has been modified to collect all of the sample from the SEC. The previous procedure collected only 100 microliter samples. The procedure was changed to insure that no components were missed. SEC's of ZAP-2, B-3, Wyodak and Big Brown product slurries are presented in this report. Quantitative analysis should be possible by the end of the next quarter. 8 refs., 15 figs., 7 tabs.

Biogas energy derived from recycled algal biomass grown on wastewater could provide a sustainable pathway for a renewable future. This research investigates the chemical details of cobalt-catalysed pyrolysis integrated with methanogenic archaea co-anaerobic fermentation to improve biogas and methane generation from wastewater algae. Algal biomass (500 mL sample) was harvested from multiple locations at the Qinghe Wastewater Treatment Plant in Beijing, China. The algal species Chlorella vulgaris and Scenedesmus obliquus were identified. A 5% Co/Al₂O₃ catalyst was prepared by impregnating commercial alumina with a cobalt nitrate solution. Pyrolysis was conducted in a 500 mL fixed-bed reactor, and bio-oil and char yields were measured. Thermal degradation of biomass and by-products was analysed using thermogravimetric analysis (TGA). Microbial cultures of Methanosaeta concilii and Methanosarcina barkeri were used for anaerobic fermentation in 1 L batch biodigesters, with bio-oil as the carbon source. Biogas production kinetics were modelled using the modified Gompertz and Arrhenius equations. Statistical analyses were performed using GraphPad Prism version 10.2.0 and R version 4.03. The results demonstrated that biogas production in the experimental group was significantly higher across all temperatures. Maximum methane yield (Pmax) increased from 301.05 mL at 400°C to 436.71 mL at 800°C in the experimental group, compared to the control group. The rate constant (k) for biogas production also increased, reaching 0.20 mL/day at 800°C in the experimental group. CO₂ yield was higher in the control group at lower temperatures, while the integrated system consistently produced more biochar and biogas. The energy efficiency analysis revealed that the calorific value of biogas increased from 7.552 MJ at 400°C to 12.966 MJ at 800°C in the experimental group, with net energy gain decreasing as temperature increased. The mass balance showed that, during the pyrolysis stage, 100 g of biomass resulted in 35 g of biochar, 250 mL of biogas, and 50 g of bio-oil. In the anaerobic digestion stage, 155.47 g of biochar and 300 mL of biogas were produced. Kinetic model analysis showed that the activation energy for pyrolysis in the experimental group decreased from 145 kJ/mol at 400°C to 125 kJ/mol at 800°C, while the maximum methane yield in the Gompertz model increased from 405.026 mL at 400°C to 434.525 mL at 800°C in the experimental group. Thermogravimetric analysis (TGA) showed that biomass had 96.8% volatile matter, while biochar had 87.5% volatile matter and 12.5% ash content. BET surface area analysis of Co/Al₂O₃ biochar showed a surface area of 400 m²/g. Cobalt-catalysed pyrolysis and the subsequent anaerobic digestion process provide synergistic effects, leading to enhanced biogas yield while reducing the production time required.

This dataset contains both 8 and 16 sampled frames of the "eating-spaghetti" video of the Kinetics-400 dataset, with the following frame indices being used:

8 frames (eating_spaghetti_8_frames.npy): 97, 98, 99, 100, 101, 102, 103, 104 16 frames (eating_spaghetti.npy): [164, 168, 172, 176, 181, 185, 189, 193, 198, 202, 206, 210, 215, 219, 223, 227]. 32 frames (eating_spaghetti_32_frames.npy): array([ 47, 51, 55… See the full description on the dataset page: https://huggingface.co/datasets/hf-internal-testing/spaghetti-video.

I3D Video Features, Labels and Splits for Multicamera Overlapping Datasets Pets-2009, HQFS and Up-Fall

The Inflated 3D (I3D) video features, ground truths, and train/test splits for the multicamera datasets Pets-2009, HQFS, and Up-Fall are available here. We relabeled two datasets (HQFS and Pets-2009) for the task of VAD-MIL under multiple cameras. Three feature dispositions of I3D data are available: I3D-RGB, I3D-OF, and the linear concatenation of these features. These datasets can be used as benchmarks for the video anomaly detection task under multiple instance learning and multiple overlapping cameras.

Preprocessed Datasets

PETS-2009 is a benchmark dataset (https://cs.binghamton.edu/~mrldata/pets2009) aggregating different scene sets with multiple overlapping camera views and distinct events involving crowds. We labeled the scenes at \textit{frame} level as anomaly or normal events. Scenes with background, people walking individually or in a crowd, and regular passing of cars are considered normal patterns. Frames with occurrences of people running (individually or in the crowd), crowding of people in the middle of the traffic intersection, and people in the counterflow were considered anomalous patterns. Videos of scenes with the occurrence of anomalous frames are labeled as anomalous, while videos without the occurrence of anomalies are marked as normal videos. The High-Quality Fall Simulation Data - HQFS dataset (https://iiw.kuleuven.be/onderzoek/advise/datasets/fall-and-adl-meta-data) is an indoor scenario with five overlapping cameras with the occurrence of fall incidents. We consider a person falling on the floor an uncommon event. We also relabeled the frame annotations to consider the intervals where the person remains lying on the ground after the fall. The multi-class Up-Fall (https://sites.google.com/up.edu.mx/har-up/) detection dataset contains two overlapping camera views and infrared sensors in a laboratory scenario.

Video Feature Extraction

We use Inflated 3D (I3D) features to represent video clips of 16 frames. We use the Video Features library (https://github.com/v-iashin/video_features) that uses a pre-trained model on the Kinetics 400 dataset. For this procedure, the frame sequence length from which to get the video clip feature representation (or window size) and the number of frames to step before extracting the next features were set to 16 frames. After the video extraction process, each video from each camera corresponds to a matrix with dimension n x 1024, where n is a variable number of existing segments and the number of attributes is 1024 (I3D attributes referring to RGB appearance information or I3D attributes referring to Optical Flow information). It is important to note that the videos (\textit{bags}) are divided into clips with a fixed number of \textit{frames}. Consequently, each video \textit{bag} contains a variable number of clips. A clip can be completely normal, completely anomalous, or mixed with normal and anomalous frames. There are three possible deep feature dispositions considered: I3D features generated with only RGB (1024 I3D features from RGB data), Optical Flow (1024 I3D features from optical flow data), and the combination of both (by simple linear concatenation). We also make available 10-crop features (https://pytorch.org/vision/main/generated/torchvision.transforms.TenCrop.html) by yielding 10 crops for a given video clip.

File Description

center-crop.zip: Folder with I3D features of Pets-2009, HQFS and Up-Fall datasets;

10-crop.zip: Folder with I3D features (10-crop) of Pets-2009, HQFS and Up-Fall datasets;

gts.zip: Folder with ground truths at frame-level and video-level of Pets-2009, HQFS and Up-Fall datasets;

splits.zip: Folder with Lists of training and test splits of Pets-2009, HQFS and Up-Fall datasets;

A portion of the preprocessed I3D feature sets was leveraged in the studies outlined in these publications:

Pereira, S. S., & Maia, J. E. B. (2024). MC-MIL: video surveillance anomaly detection with multi-instance learning and multiple overlapped cameras. Neural Computing and Applications, 36(18), 10527-10543. Available at https://link.springer.com/article/10.1007/s00521-024-09611-3.

Pereira, S. S. L., Maia, J. E. B., & Proença, H. (2024, September). Video Anomaly Detection in Overlapping Data: The More Cameras, the Better?. In 2024 IEEE International Joint Conference on Biometrics (IJCB) (pp. 1-10). IEEE. Available at https://ieeexplore.ieee.org/document/10744502.

This repository contains the mapping from integer id's to actual label names (in HuggingFace Transformers typically called id2label) for several datasets. Current datasets include:

ImageNet-1k ImageNet-22k (also called ImageNet-21k as there are 21,843 classes) COCO detection 2017 COCO panoptic 2017 ADE20k (actually, the MIT Scene Parsing benchmark, which is a subset of ADE20k) Cityscapes VQAv2 Kinetics-700 RVL-CDIP PASCAL VOC Kinetics-400 ...

You can read in a label file as follows (using… See the full description on the dataset page: https://huggingface.co/datasets/huggingface/label-files.

Tetralin is the simplest polycyclic naphtheno-aromatic hydrocarbon found in liquid fuels (e.g., jet fuels, diesel). It is also produced by the pyrolysis and oxidation of decalin. To get a better understanding of tetralin combustion characteristics, new oxidation experiments were performed in a jet-stirred reactor, and the results are reported here. For the first time, stable species concentration profiles were measured at 1 and 10 atm over a range of equivalence ratios (φ = 0.5–1.5) and temperatures (790–1400 K). The oxidation of tetralin under these conditions was simulated using a semidetailed kinetic reaction scheme (∼10 000 reactions and ∼400 species) deriving from a chemical kinetic model proposed earlier for the oxidation of decalin over a wide range of conditions (jet-stirred reactor, plug-flow reactor, and shock-tube). The proposed kinetic model shows reasonable agreement with the present measurements. It can also be used to represent tetralin pyrolysis based on a variety of results available in the literature. Sensitivity analyses and reaction pathway computations were used for rationalizing the results.