Multimodal

Search our off-the-shelf datasets.

Filter by
Language
Filter by Languages
Language
Devices
Devices
Applicable Fields
Applicable Fields
More
Applicable Scenarios
Applicable Scenarios
2D Flat Anti‑Spoofing Attack Video Dataset
This dataset collects live attack video data of Chinese people. The collection scenarios cover indoor and outdoor environments and various light conditions. The shooting distance will be dynamically adjusted according to the actual shooting situation, mainly changing between full-body shots and head-and-shoulder shots. The types of attack methods collected include static attacks and dynamic attacks, totaling 7 categories. Static attack forms include photos, A4 color prints, and black-and-white prints. Dynamic attack forms include playing videos on iPads, Android phones, iPhones, and laptops. Each person will shoot 6 live videos and 22 attack videos in a normal and natural state, with each video lasting approximately 15 seconds. The collection tools are collected using a handheld method.
3D modeling data or market sence
The data collection equipment for this dataset is a 3D space scanning device. It was collected in Chinese supermarkets. Using a 3D space scanning camera device, 10 complete 3D scene data of medium and small-sized supermarkets were collected, and a 3D point cloud data model was generated for each supermarket.
Analog brake video dataset
This dataset collects video data of the personnel passing through the gate. The collection scenario is indoors and all the shooting is under normal lighting conditions. The installation of the cameras is based on the height and angle of the simulated gate passage. People are queued up for passage in groups of 1 to 3 (no backside data is collected). Each collector will take a frontal head-and-shoulders photo with a camera and record 12 random number audio clips with a mobile phone.
Conference video action collection dataset
This dataset is a simulated video dataset of meeting actions. It captures various skin colors including dark, brown, fair, and light, recorded in a bright meeting room environment, with a single person's frontal video, using 4 different collection devices. Each person has 4 video clips and 2 pictures collected.
Ego-centric High-quality Dataset
This dataset covers more than 10 types of scenarios, including residential settings, offices, retail stores, restaurants, manufacturing facilities, warehouses, agricultural environments, and entertainment venues. It includes over 200 tasks and provides rich first-person action and human–environment interaction data. The videos have a resolution of 1920 × 1080 and a frame rate of 60 fps. The dataset can be used for embodied AI model training, egocentric behavior understanding, environmental perception and interaction, and other related applications.
Figure reference reading dataset
This dataset consists of video recordings of finger-pointing actions. The videos capture the complete movement of the index finger pointing from outside the camera lens towards the printed English words on a book. The finger must remain on the book for at least 2 seconds. The light sources used during the collection include LED lights and incandescent lamps, and the light coverage includes strong, medium, and weak types. The camera captures both the left and right hands. The captured range of the fingers includes the entire hand and arm, the entire index finger, half of the index finger, and four types: the entire hand, the entire index finger, half of the index finger, and the entire hand plus the arm. The types of books also include those with doodles and clean books, with or without creases, with or without curling. The collection equipment covers 5 positions: the front, upper right, upper left, right side, and left side of the screen, and is shot in landscape mode. The main subjects of the collection are children and middle-aged people.
Hospital Multi‑Scene Patient Action Simulation Video Dataset I
This dataset consists of video data collected in a simulated hospital environment. The collectors captured the three postures of the subjects: lying on the bed, sitting on the bed, and standing outside the bed. A total of 118 individuals were included, and each person had 3 video segments collected. The resolution of the videos included 720P and 1080P.
Hospital Multi‑Scene Patient Action Simulation Video Dataset II
This dataset consists of video data collected in a simulated hospital environment, covering approximately 10 different scenarios. It includes shots of at least three types of different hospital beds. The subjects are shown in three postures: lying on the bed, sitting on the bed, and standing outside the bed. Each model is filmed three times. The first time, they are filmed in their own clothes. The second time, they change into hospital gowns (usually white or blue) for the shot. The third time, they can freely express themselves in the hospital gowns.
Lip speech video was collected for 250 people
This dataset uses six cameras and two microphones to simultaneously collect and record the lip-voiced video data of the speaker. The shooting scene is simulated in a cockpit environment, and the shooting angles and lighting conditions are diverse.

Join our newsletter to stay updated

Thank you for signing up!

Stay informed and ahead with the latest updates, insights, and exclusive content delivered straight to your inbox.

By subscribing you agree to with our Privacy Policy and provide consent to receive updates from our company.

Filter by
Filter by
Language
Filter by Languages
Language
Devices
Devices
Applicable Fields
Applicable Fields
More
Applicable Scenarios
Applicable Scenarios