TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Applied ML: Composition-Aware Search for Shutterstock Photography

22 点作者 kscz超过 7 年前

6 条评论

mranzinger超过 7 年前
As the primary researcher and engineer on this project, I'm really excited to finally have been able to release this, and it's also the first system we've shown off that includes a technical whitepaper. Nothing would make me happier than seeing a paper come out in Arxiv that improves this system.
bencaggia超过 7 年前
This is a great application of computer vision applied to everyday business problems. Kudos to the shutterstock search/CV team.
kscz超过 7 年前
The whitepaper linked in the upper-right is highly informative: <a href="https:&#x2F;&#x2F;www.shutterstock.com&#x2F;labs&#x2F;compositionsearch&#x2F;static&#x2F;cas-final.pdf" rel="nofollow">https:&#x2F;&#x2F;www.shutterstock.com&#x2F;labs&#x2F;compositionsearch&#x2F;static&#x2F;c...</a>
STHayden超过 7 年前
Finding good copy space is such a time consuming part of looking for images and any way to make that easier would be welcome!
niamh超过 7 年前
Love the ability to find images with copy space - I&#x27;ve never seen that before. It&#x27;s also really fast.
siobhanaa超过 7 年前
Truly unique tool. Will save designers hours.