18,880 Images of 466 People - 3D Instance Segmentation and 22 Landmarks Annotation Data of Human Body

Human body instance segmentation

Human body landmarks

Multiple poses

Multiple scenes

18,880 Images of 466 People - 3D Instance Segmentation and 22 Landmarks Annotation Data of Human Body. The dataset diversity includes multiple scenes, light conditions, ages, shooting angles, and poses. In terms of annotation, we adpoted instance segmentation annotations on human body. 22 landmarks were also annotated for each human body. The dataset can be used for tasks such as human body instance segmentation and human behavior recognition.

This is a paid datasets for commercial use, research purpose and more. Licensed ready made datasets help jump-start AI projects.

Specifications

Data size

466 people, the number of images is 18,880

Race distribution

Asian

Gender distribution

233 males, 233 females

Age distribution

299 children and teenagers, 167 adults

Collecting environment

including indoor scenes and outdoor scenes

Data diversity

multiple scenes, light conditions, ages, shooting angles, and poses

Device

RealSense Depth Camera D435i

Data format

the image data format is .jpg and .png, the annotation file format are .json，the file format of camera parameters is .txt

Collecting content

3D images of various human body poses ((the rgb channel and depth channel have been registered))

Annotation content

instance segmentation annotation of human body , 22 landmarks annotation of human body

Accuracy

Accuracy requirement: the mask edge location errors in x and y directions are less than 3 pixels, which is considered as a qualified annotation; Accuracy requirement of segmentation annotation: the annotation part (each part of mask) is regarded as the unit, the accuracy rate shall be more than 95%; Accuracy requirement of landmark annotation: the annotation part (each landmark) is regarded as the unit, the accuracy rate shall be more than 95%; Annotating accuracy = number of correct annotations / total number of annotations

Recommended Dataset

15 People - 22 Landmarks Annotation Data of 3D Human Body

15 People - 22 Landmarks Annotation Data of 3D Human Body. The dataset diversity includes multiple scenes, different ages, different costumes, different human body sitting postures. In terms of annotation, we annotate the 2D and 3D coordinates of the 22 landmarks of the human body, landmark attributes, the rectangular frame of the human body. The dataset can be used for tasks such as human body instance segmentation and human behavior recognition.

Human body instance segmentation Human body landmarks 3D Multiple poses Multiple scenes

25,080 Images – Alpha Matte Human Body Segmentation Data (Basic Version)

25,080 Images – Alpha Matte Human Body Segmentation Data. The data includes indoor scenes and outdoor scenes. The dataset diversity includes multiple scenes, multiple age groups, multiple human body angles, multiple postures. In terms of annotation, alpha matte segmentation annotation was adopted for the human body. The data can be used for tasks such as alpha matte human body segmentation.

Alpha matte human body segmentation multiple scenes multiple age groups multiple human body angles multiple postures

602 People –3,010 Images Multi-Races Human Body Semantic Segmentation Data

602 People –3,010 Images Multi-Races Human Body Semantic Segmentation Data,The data diversity includes headphones, body, background,and glasses.In terms of annotation, we adpoted segmentation annotations on headphones, body, background and glasses.The data can be used for tasks such as human body segmentation and the behavior detection of Video conference.

Human body segmentation different poses different ages different races different collection backgrounds different scenes.

21,300 Images - Human Body Segmentation Data

21,300 Images - Human Body Segmentation Data. The data includes indoor scenes and outdoor scenes. The data covers female people and male people. The race distribution includes Asian, black race and Caucasian. The age distribution ranges from teenager to the elderly, the middle-aged and young people are the majorities. The dataset diversity includes multiple scenes, ages, races, postures, and appendages. In terms of annotation, we adpoted pixel-wise segmentation annotations on human body. The data can be used for tasks such as human body segmentation.

Multiple scenes Multiple ages Multiple races Multiple postures Multiple appendages

34,981 Images – Alpha Matte Human Body Segmentation Data (Fine Version)

34,981 Images – Alpha Matte Human Body Segmentation Data. The data includes indoor scenes and outdoor scenes. The dataset diversity includes multiple scenes, multiple age groups, multiple human body angles, multiple postures. In terms of annotation, alpha matte segmentation annotation was adopted for the human body. The data can be used for tasks such as alpha matte human body segmentation.

Alpha matte human body segmentation multiple scenes multiple age groups multiple human body angles multiple postures

144,810 Images Multi-class Fashion Item Detection Data

144,810 Images Multi-class Fashion Item Detection Data. In this dataset, 19,968 images of male and 124,842 images of female were included. The Fashion Items were divided into 4 parts based on the season (spring, autumn, summer and winter). In terms of annotation, rectangular bounding boxes were adopted to annotate fashion items. The data can be used for tasks such as fashion items detection, fashion recommendation and other tasks.

Fashion Item Detection Multiple scenes Different seasons Different races

49,945 Images Human Costume & Apparel Accessory Segmentation Data

49,945 Images Human Costume & Apparel Accessory Segmentation Data. The gender distribution includes female and male, the race distribution is Asian, Caucasian and black race, the age distribution is teenager, young and middle-aged. The data diversity includes multiple scenes, multiple light conditions, multiple types of costume (upper garment, lower garment, and shoes), and multiple apparel accessories (bag, glasses, accessories, etc.). In terms of annotation, semantic segmentation of 47 categories object (including background, costume and apparel accessory) was adopted. The dataset can be used for tasks such as human costume & apparel accessory segmentation and fashion recommendation.

Human costume & apparel accessory segmentation multiple scenes multiple light conditions multiple types of costume multiple apparel accessories

50,356 Images - Human Body Segmentation and 18 Landmarks Data

50,356 Images - Human Body Segmentation and 18 Landmarks Data. The data diversity includes multiple scenes, ages, races, poses, and appendages. In terms of annotation, we adpoted segmentation annotations on human body and appendages.18 landmarks were also annotated for each human body. The data can be used for tasks such as human body segmentation and human behavior recognition.

Human body segmentation Landmark Multiple races Multiples postures Multiple scenes