Dataset for human portrait segmentation, about 15k images in 128x128 size

Context

I picked up this dataset from https://github.com/anilsathyan7/Portrait-Segmentation. Since they did not specify the license for the dataset, I will assume that the data set is the same license as the repository (MIT). This dataset is processed, while the original comes in npy format.

Content

There are folders, the one starts with x contains the image, the one start with y contain the mask of the human face. train and test is self-explainatory. The mask images have one channel only.

Portrait segmentation is the process whereby the head and upper body of a person is separated from the background of an image or video stream. This is difficult to achieve accurately, although good results have been obtained with deep learning methods which cope well with occlusion, pose and illumination changes. These are however either slow or require a powerful system to operate in real-time. We present a new method of portrait segmentation called FaceSeg which uses fast DBSCAN clustering combined with smart face tracking that can replicate the benefits and accuracy of deep learning methods at a much faster speed. In a direct comparison using a standard testing suite, our method achieved a segmentation speed of 150 fps for a 640x480 video stream with median accuracy and F1 scores of 99.96% and 99.93% respectively on simple backgrounds, with 98.81% and 98.13% on complex backgrounds. The state-of-art deep learning based FastPortrait / Mobile Neural Network method achieved 15 fps with 99.95% accuracy and 99.91% F1 score on simple backgrounds, and 99.01% accuracy and 98.43 F1 score on complex backgrounds. An efficacy-boosted implementation for FaceSeg can achieve 75 fps with 99.23% accuracy and 98.79% F1 score on complex backgrounds.

Dataset

This dataset was created by Ajay Vamsi

Contents

Dataset Card for AISegment.cn - Matting Human datasets

  Dataset Description

Quoting the dataset's github (translated by Apple Translator):

This dataset is currently the largest portrait matting dataset, containing 34,427 images and corresponding matting results. The data set was marked by the high quality of Beijing Play Star Convergence Technology Co. Ltd., and the portrait soft segmentation model trained using this data set has been commercialized.

The… See the full description on the dataset page: https://huggingface.co/datasets/fredguth/aisegmentcn-matting-human.

P3M-10k , which is the first large-scale anonymized benchmark for Privacy-Preserving Portrait Matting. P3M-10k consists of 10,000 high-resolution face-blurred portrait images along with high-quality alpha mattes. We systematically evaluate both trimap-free and trimap-based matting methods on P3M-10k and find that existing matting methods show different generalization capabilities when following the Privacy-Preserving Training (PPT) setting, ?.?., training on face-blurred images and testing on arbitrary images.

Face segmentation

An example of a dataset that we've collected for a photo edit App. The dataset includes 20 selfies of people (man and women) in segmentation masks and their visualisations.

💴 For Commercial Usage: To discuss your requirements, learn about the price and buy the dataset, leave a request on TrainingData to buy the dataset

SIMILAR DATASETS:

• Bald People Segmentation Dataset
• Dataset of bald people, 5000 images
• Body Segmentation - 5,300 Photos
• Selfies, ID Images dataset (5591 sets of 15 files)
• Hair Detection & Segmentation Dataset

Content

There are two folders: - img - original images of faces - masks - segmentation masks for faces

On your demand we collect similar dataset according to your technical instructions (guides) to help you with your specific project's targets.

💴 Buy the Dataset: This is just an example of the data. Leave a request on https://trainingdata.pro/data-market to discuss your requirements, learn about the price and buy the dataset

TrainingData provides high-quality data annotation tailored to your needs.

keywords: head segmentation dataset, face-generation, semantic segmentation, face parts recognition, human faces, portrait segmentation, human face extraction, image segmentation, annotation, biometric dataset, biometric data dataset, face recognition database, facial recognition, face forgery detection, face shape, facial gestures, ar, augmented reality, face recognition dataset, face detection dataset, facial analysis, human images dataset

Our "Single-person Portrait Matting Dataset" is a pivotal resource for the fashion, media, and social media industries, providing finely labeled portrait images that capture a wide range of postures and hairstyles from various countries. With a focus on high-resolution images exceeding 1080 x 1080 pixels, this dataset is tailored for applications requiring detailed segmentation, including hair, ears, fingers, and other intricate portrait features.

Our "Eastern Asia Single-person Portrait Matting Dataset" targets the nuanced requirements of the fashion, internet, and entertainment sectors, featuring single-person portraits from Eastern Asia in a variety of settings including indoor, outdoor, street, and sport. This dataset is specially curated for pixel-level fine segmentation tasks, capturing diverse postures and scenarios.

A large-scale video portrait dataset that contains 291 videos from 23 conference scenes with 14K frames. This dataset contains various teleconferencing scenes, various actions of the participants, interference of passers-by and illumination change.

Snapshot img

Global Digital Photo Frame (DPF) Market Forecast 2023-2027

The global digital photo frame (DPF) market is forecast to increase by USD 110.9 million and is estimated to grow at a CAGR of 3.03% between 2022 and 2027. The growth trajectory of the global digital photo frame market is shaped by several key factors. These include the rising demand for enhanced living standards, the widespread adoption of smart and connected home systems, and ongoing product innovations driving premiumization within the industry. As consumers increasingly seek sophisticated solutions for displaying and sharing digital memories, the market experiences a surge in demand. Moreover, the integration of advanced features and connectivity options in digital photo frames caters to the evolving needs of tech-savvy consumers, driving market expansion further. This report includes an in-depth analysis of drivers, trends, and challenges. Furthermore, the report includes historic market data from 2017 to 2021.

What will be the size of the Digital Photo Frame Market During the Forecast Period?

To learn more about this report, Download Report Sample

Market Segmentation

The offline segment is estimated to witness significant growth during the forecast period. The offline distribution channel comprises specialty stores, department stores, hypermarkets, supermarkets, convenience stores, and clubhouse stores and department stores.

Get a glance at the market contribution of various segments Request a PDF Sample

The offline segment was the largest segment and was valued at USD 562.60 million in 2017. In this segment, the specialty stores comprise traditional furniture/furnishing/home decor stores, home improvement stores, consumer electronics stores, lifestyle stores, home accents/gift stores, and craft/fabric stores and chains. Some hypermarkets offering digital picture frames are expanding their presence as well.

Key Regions

For more insights on the market share of various regions Request PDF Sample now!

North America is estimated to contribute 37% to the growth of the global market during the forecast period. Technavio’s analysts have elaborately explained the regional trends and drivers that shape the market during the forecast period.

The main factors that contribute to the growth of the market in the US are the launch of innovative products, the growing use of online platforms to shop for home decor products such as digital photo frames, and the rising adoption of technologically advanced products by consumers The North American market is also being stimulated by the increase in demand for personalized digital photo frames. Even though the saturating economic condition of the US is expected to hamper the growth of the market during the forecast period, the surging demand for these frames will drive the growth of the market in North America during the forecast period.

Market Dynamics and Customer Landscape

The Global Digital Photo Frame Market offers a wide array of options, catering to customers' diverse preferences, with various display sizes ranging from small to large, providing a perfect fit for every budget and disposable income, equipped with features such as built-in speakers and support for digital images, video files, and popular formats like M-JPEG, WMA, AVI, MPEG-1, MPEG-2, and MPEG-4, ensuring an immersive viewing experience for cherished memories in portrait orientation. Our researchers analyzed the data with 2022 as the base year, along with the key drivers, trends, and challenges. A holistic analysis of drivers will help companies refine their marketing strategies to gain a competitive advantage.

Key Market Driver

The inclination toward a high standard of living is the key factor driving global market growth. The market experiences a boost as residential renovations become more prevalent, driven by enhanced lifestyles and higher disposable incomes. Consumers seek modern home decor solutions like digital picture frames to enhance their living spaces, catering to their preference for personalized products and the availability of a wide range of display sizes, from small to large.

This trend is further accelerated by the growing influence of the Internet and social media, providing consumers with insights into the latest product trends and advancements in digitized photo frames. With features such as built-in speakers and support for various video file formats like M-JPEG, WMA, AVI, MPEG-1, MPEG-2, and MPEG-4, digital photo frames offer an immersive viewing experience for cherished memories captured in portrait orientation.

Significant Market Trends

The introduction of smart digital photo frames is the primary trend in the global market. The growth opportunity for smart digital photo frames has been stimulated by a strong rise in demand for home automation and Smart Hubs. Repeat purchases hold a major share in the sales of smart digital ph

The "Facial 17 Parts Segmentation Dataset" is specifically compiled for the visual entertainment industry, featuring a range of internet-collected facial images with resolutions exceeding 1024 x 682 pixels. This dataset is dedicated to semantic segmentation, delineating 17 facial categories such as eyebrows, lips, eye pupils, and more. It also includes a selection of portrait images with occlusions, adding complexity and diversity to the dataset for more realistic application scenarios.

Description： 5,011 Images – Human Frontal face Data (Male). The data diversity includes multiple scenes, multiple ages and multiple races. This dataset includes 2,004 Caucasians , 3,007 Asians. This dataset can be used for tasks such as face detection, race detection, age detection, beard category classification.

Data size： 5,011 people, one image per person

Race distribution： 2,004 Caucasians , 3,007 Asians

Synthetic Faces High Quality (SFHQ) part 2

This dataset consists of 91,361 high quality 1024x1024 curated face images, and was created by "bringing to life" various 3D models and correcting bad "text to image" generations from stable diffusion model using a process similar to what is described in this short twitter thread which involve encoding the images into StyleGAN2 latent space and performing a small manipulation that turns each image into a photo-realistic image.

The dataset also contains facial landmarks (extended set) and face parsing semantic segmentation maps. An example script is provided and demonstrates how to access landmarks, segmentation maps, and textually search withing the dataset (with CLIP image/text feature vectors), and also performs some exploratory analysis of the dataset. link to github repo of the dataset.

The process that "brings to life" face-like images and creates several candidate photo-realistic ones is illustrated below: https://i.ibb.co/0sw8TkL/bring-to-life-process-SD-FS-2.png" alt="">

More Details

1. The original inspiration images are taken from Face Synthetics Dataset which contains 3D models of faces and generated images using stable diffusion v1.4 model using various face portrait prompts that span a wide range of ethnicities, ages, expressions, hairstyles, etc. Note that stable diffusion faces often contain extreme error in the generation (as one can be seen in the three rightmost columns in the image above) so cannot be used to create a photo-reallistic dataset without a correcting model or an extremely lengthy manual curation process.
2. Each inspiration image was encoded by encoder4editing (e4e) into StyleGAN2 latent space (StyleGAN2 is a generative face model tained on FFHQ dataset) and multiple candidate images were generated from each inspiration image
3. These candidate images were then further curated and verified as being photo-realistic and high quality by a single human (me) and a machine learning assistant model that was trained to approximate my own human judgments and helped me scale myself to asses the quality of all images in the dataset
4. Near duplicates and images that were too similar were removed using CLIP features (no two images in the dataset have CLIP similarity score of greater than ~0.92)
5. From each image various pre-trained features were extracted and provided here for convenience, in particular CLIP features for fast textual query of the dataset
6. From each image, semantic segmentation maps were extracted using Face Parsing BiSeNet and are provided in the dataset under "segmentations"
7. From each image, an extended landmark set was extracted that also contain inner and outer hairlines (these are unique landmarks that are usually not extracted by other algorithms). These landmarks were extracted using Dlib, Face Alignment and some post processing of Face Parsing BiSeNet and are provided in the dataset under "landmarks"
8. NOTE: semantic segmentation and landmarks were first calculated on scaled down version of 256x256 images, and then upscaled to 1024x1024

Parts 1,2,3,4

• Part 1 of the dataset consists of 89,785 HQ 1024x1024 curated face images. It uses "inspiration" images from Artstation-Artistic-face-HQ dataset (AAHQ), Close-Up Humans dataset and UIBVFED dataset.
• Part 2 of the dataset consists of 91,361 HQ 1024x1024 curated face images. It uses "inspiration" images from Face Synthetics dataset and by sampling from the Stable Diffusion v1.4 text to image generator using varied face portrait prompts.
• Part 3 of the dataset consists of 118,358 HQ 1024x1024 curated face images. It uses "inspiration" images by sampling from StyleGAN2 mapping network with very high truncation psi coefficients to increase diversity of the generation. Here, the e4e encoder is basically used a new kind of truncation trick.
• Part 4 of the dataset consists of 125,754 HQ 1024x1024 curated face images. It uses "inspiration" images by sampling from the Stable Diffusion v2.1 text to image generator using varied face portrait prompts.
• See also dataset github repo with full details and links

Summary

Overall, the SFHQ dataset contains ~425,000 high quality and curated synthetic face images that have no privacy issues or license issues surrounding them.

This dataset contains a high degree of variability on the axes of identity, ethnicity, age, pose, expression, lighting conditions, hair-style, hair-color, facial hair. It lacks variability in accessories axes such as hats or earphones as well as various jewelry. It also doesn't contain any occlusions except the self-occlusion of hair occluding the forehead, the ears and rarely the eyes. This dataset naturally inherits all the biases of it's original datasets (FFHQ, AAHQ, Close-Up Humans, Face Synthetics, LAION-5B) and the StyleGAN2 and Stable Diffusion models.

The purpose of this dataset is to be of sufficiently high quality that new machine learning models can be trained using this data, including even generative face models such as StyleGAN. The dataset may be extended from time to time with additional supervision labels (e.g. text descriptions), but no promises.

Hope this is helpful to some of you, feel free to use as you see fit...