New dataset and deep learning model estimate human weight and height from images

Researchers have developed a method for estimating human weight and height from single images captured in everyday settings. This approach utilizes deep neural networks and explores various data modalities, including RGB, depth maps, and pose-affinity maps, to predict Body Mass Index (BMI), weight, and height. To facilitate this research, a new dataset of 6,105 images with ground truth labels for these metrics has been created, featuring diverse poses, backgrounds, and image qualities. Experiments using CNN backbones like VGG, DenseNet, and ResNet demonstrated that full-body images yield superior results compared to half-body or facial images. AI

IMPACT This research could lead to new applications in health monitoring, fitness tracking, and personalized recommendations based on visual data.

RANK_REASON This is a research paper detailing a new method and dataset for estimating human physical characteristics from images. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• CatalyzeX
• Connected Papers
• CORE Recommender
• DagsHub
• DenseNet
• Gotit.pub
• Hugging Face
• Litmaps
• residual neural network
• ScienceCast
• scite Smart Citations
• Vgg Neural Network

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →