A few months ago, Google released a program called 'Quick, Draw!' that trained a neural network to recognize human drawings through the guise of a fun interactive game, and now Google has updated the game's website to 'Quick, Draw! The Data' in order to open source some of the company's findings.
The Quick, Draw! game was simple on its face. The site would post a random word (usually a concrete object or simple concept,) and users had 20 seconds to draw that word. The aim of the game was to get Google's neural network to recognize the image within the time limit. Quick, Draw! went viral, and Google subsequently collected millions of drawings of hundreds of words.
Quick, Draw! The Data lays out all those drawings, organized by word. Clicking through, users can see a sped up animation of the exact lines that people drew as they played, giving a thorough and enlightening picture of how people interpret words and concepts visually.
Animated Drawing Datasets
'Quick Draw! The Data' Shows How People Drew the Same Concepts
Trend Themes
1. Interactive Drawing - Using interactive drawing games to collect data for machine learning can lead to more accurate recognition and insights.
2. Open Sourcing Data - Companies can benefit from open sourcing their data by gaining community insights and accelerating innovation.
3. Visual Interpretation - Analyzing people's visual interpretation of words and concepts can reveal valuable insights for design and marketing.
Industry Implications
1. Education Technology - Interactive drawing games can be incorporated into educational technology to improve learning outcomes through visual learning.
2. Advertising - Visual interpretation data can be harnessed by advertising firms to create more effective and engaging advertisements.
3. Artificial Intelligence - Open sourcing drawing datasets can facilitate the development of more accurate and sophisticated machine learning algorithms.