Free Sample Dataset - 1000 High Resolution Images & Metadata|图像数据数据集|机器学习数据集

**Overview**
This is a free sample dataset consisting of 1000 images and accompanying metadata sourced from our +550 million image library. Image types for this sample include photos, vectors, and illustrations across a vast range of content categories and settings. This sample includes a wide range of metadata fields including content descriptions and keywords that make it ideal for powering a wide variety of machine learning use cases.

If you’d like to start licensing data from the full range of imagery and metadata available at Shutterstock please reach out directly to our team at [e-mail] to start using our tailored services to help ideate, curate and customize datasets for your unique business needs.

**Use cases**
This type of data can be licensed from Shutterstock for a wide variety of use cases including powering machine learning models that have generative capabilities.

**Metadata**

Sample metadata fields included in this dataset are listed below, for a full list of all metadata available from Shutterstock please contact our team at [e-mail].

**asset metadata:**

id
keywords
image_type
is_creative
mature_flag
date_submitted
date_captured
asset_location
popularity_score
german_description
spanish_description
french_description
korean_description
japanese_description
labels
moderation_labels
has_model_release
has_people
primary_category


**file metadata:**

asset_id
asset_file_size
asset_file_extension
asset_file_size_in_bytes
width
height
orientation


**model metadata:**
asset_id
model_release_id
age_range
age_in_years
gender
ethnicity

**Our data**

Shutterstock offers the largest, highest quality and most diverse collection of creative content with best-in-class metadata, giving technology businesses the scale and accuracy they need to build and sustain a wide variety of machine learning models. Our growing library of +550M images, +40M videos, +4M music and audio tracks, and +1.2M 3D models and data is human-reviewed for accuracy and IP infringement, allowing you to use our data worry-free and avoid unwanted or unlawful content. We ethically source all content from over 2 million creators in +150 countries and with over 60 million new assets added annually, our ever-growing library gives you access to fresh and diverse datasets that can be refreshed regularly to meet all your data needs.