Computer Vision Datasets

2,255 Images – Multi-Race Human Body Semantic Segmentation Dataset

This dataset includes 2,255 images of 451 people across multiple races. The semantic segmentation area includes headphones, glasses, body and background.This dataset can be used for training AI models in human body segmentation, video conference behavior detection, and smart vision applications.

human body segmentation dataset semantic segmentation dataset human multi-race human dataset human parsing dataset video conference dataset body part segmentation dataset human behavior recognition dataset annotated human dataset

31.5K Images – 106 Facial Landmark Detection Dataset

This dataset contains 31,549 facial images annotated with 106 facial landmarks. It includes Black, White, and Indian individuals.The dataset covers diverse and challenging conditions, including multiple scenes, facial poses, age groups, lighting conditions, and complex facial expressions. This data can be used for facial landmark detection.

facial landmark dataset facial landmark detection dataset facial landmark annotation dataset face landmark dataset facial keypoint dataset

900-Person Multi-Ethnic Facial Video Dataset – Face Recognition & Liveness Detection

This dataset contains facial video data from 900 participants across multiple ethnic backgrounds and languages. Each participant is associated with two video clips and one TXT document. One video records the participant reading the provided document aloud with natural lip movements and facial expressions, while the other captures the participant remaining completely silent with a natutural lip movements and facial expressions, while the other captures the participant remaining completely silent with a natural facial expression. Participants are instructed to simulate a video conference scenario, looking directly at the camera while framing their upper body from the waist or chest upward. The dataset features diverse acquisition conditions, including multiple ethnicities, languages, age groups, document topics, backgrounds. The data can be used for application scenarios such as cross-ethnic face recognition, liveness detection, digital humans, and video face swapping.

facial video dataset face video dataset facial recognition dataset face recognition video dataset multi ethnic facial dataset

442 Images - IR Sensing Images(Airport & Port)

A total of 442 sets of infrared remote sensing image data are provided. Each set includes the original remote sensing images, single-channel infrared remote sensing images, three-channel visible light remote sensing images, fused remote sensing images, target detection annotation files, semantic segmentation annotation files, and semantic segmentation mask images. These can be used for tasks such as target detection and semantic segmentation of remote sensing images.

Infrared remote sensing Object detection Semantic segmentation Airport Port

Airport Remote Sensing Dataset for Object Detection & Segmentation – 1,437 Images

This dataset contains 1,437 remote sensing images covering airport and port scenes across different countries and geographic locations. The diverse imagery represents various real-world airports and ports. The dataset is suitable for object detection and semantic segmentation in remote sensing imagery.

remote sensing dataset remote sensing images dataset remote sensing object detection dataset remote sensing segmentation dataset

3,919 People Multi-pose Faces Dataset

3,919 People Multi-pose Faces Data, 24 images and 9 videos per person. The collection environment includes indoor and outdoor scenes. This data can be used for face detection, face recognition and other tasks.

Different face poses Nationalities Ages Light conditions Different scenes

16,270 Videos - Passenger Lose Behavior Data

16,270 Videos - Passenger Lose Behavior Data, covering loss events such as item loss and pet loss. It can be used for tasks including lost‑event detection.

Passenger lost Visible light Infrared In-car scene

38,611 Images – Human Face Segmentation Data

Human Face Segmentation Data from 38,611 Images. Pure color backgrounds, interior and exterior scene types are all included in the data. Both males and females are included in the data. Black and Caucasian races are represented in the race distribution. The data covers multiple age groups. Simple and complex facial expressions can be found in the data (large-angle tilt of face, closing eye, glower, pucker, opening mouth, etc.). We used pixel-by-pixel segmentation annotations to annotate the human face, the five sense organs, the body, and appendages. The information can be applied to tasks like facial Recon Related Tasks.

Human face segmentation Five sense organs segmentation Multiple races Multiple ages Multiple scenes Multiple expressions Multiple appendages

8.3K Images – Facial Landmark Dataset for Face Recognition

This dataset contains 8,308 facial images annotated with 106 facial landmarks. It includes Black and Caucasian individuals. In this dataset, most images feature frontal faces, while additional samples capture exaggerated facial expressions. The dataset can be used for facial landmark detection and related computer vision tasks, including face detection, face recognition, face segmentation and other tasks.

facial landmark detection dataset facial landmark dataset facial landmark annotation dataset face landmark dataset facial keypoint dataset