使用集合让一切井井有条
根据您的偏好保存内容并对其进行分类。
eBay Motors 使用 Firebase ML 快速对图片进行分类、降低成本并改善用户体验
简介
eBay Motors 让
用户能够搜寻其所在地区内待售的车辆。用户还可以通过移动应用，
在一个方便易用的界面中
上传自己想要销售的车辆的照片。
这款应用面向汽车爱好者，
也就是真正关注汽车的客户，
在他们心目中，汽车绝不仅仅是一种代步工具。
挑战
这类买家希望汽车具有更详细的描述，
也希望能更深入地了解车辆使用史。对于卖家来说，
在应用中为其车辆商品详情上传出色的照片非常重要。
但要为每张照片添加
"外饰"、"内饰"或"发动机"这类标签需要耗费大量人力，
这可能会延误待售车辆的上架，
最终使更多人不愿
通过该应用上架自己的车辆。
eBay Motors 的应用开发团队付出了很多时间
去尝试寻找自动为这些照片添加标签的最佳方法，
但一直没找到满意的解决方案。
解决方案
该团队意识到，他们并不需要在内部自行构建所有这些功能，
于是选择了借助 AutoML Vision Edge
来帮助他们应对这一用户体验挑战。Bay Motors 的原生应用负责人 Jack Hall 说：
"这里涉及到两个关键考虑事项：我们该如何为卖家提供无缝的上传体验，
减少他们在创建商品详情时遇到的阻力；
以及我们该如何降低支持这类功能的成本，
无论是工程工作成本，
还是服务器相关成本。"
AutoML 直接解决了这两个问题，
"它使我们能够缩短
通过应用创建商品详情所需的时间，
这显然对我们的卖家投资回报率有显著影响，"Hall 补充说。
借助 AutoML Vision Edge，
eBay Motors 团队使用自己的数据集在一周内就完成了
从概念到原型的过程。在执行数小时的数字标签添加任务之后，
他们就将模型改进到了
能放心地在生产环境中使用的程度。
成果
1 周时间完成从概念到原型的过程（使用他们自己的数据集）
只需执行 20 小时的数据标签添加任务即获得可在生产中使用的模型
"AutoML 帮助我们缩短了通过应用创建列表所需的时间，这对我们的卖家投资回报率产生了明确的积极影响。"
- eBay Motors 原生应用主管 Jake Hall
[null,null,[],[],[],null,["[*arrow_back*\nExplore other case studies](/case-studies) \n[View more solutions](/solutions) \n\neBay Motors uses Firebase ML to quickly categorize images, reduce costs and improve user experience \n\n*** ** * ** ***\n\nIntroduction [eBay Motors](https://pages.ebay.com/motors/motorsapp/) allows\nusers to search and find cars for sale in their area. Users can also\nupload pictures of cars that they want to sell, using a simple\nand easy-to-use inteace through the mobile app. The app is\ntargeted towards auto enthusiasts: customers who care about cars.\nThese are folks that see a car as something much more than a\nutility that helps them get from point A to B.\n\n*** ** * ** ***\n\nChallenge This type of buyer expects a more detailed description of the car\nand insights into its history. For sellers, it's of the utmost\nimportance to have well-taken pictures uploaded for their car\nlistings in the app. However, tagging each photo as an exterior,\ninterior or engine requires a lot of manual effort, which can delay\nposting a car for sale, and ultimately prevent more people from\nlisting their cars through the app.\n\nThe eBay Motors app development team spent a lot of cycles trying\nto figure out how best to automate the tagging of these photos, but\nstill didn't have a solution they were happy with. \n\n*** ** * ** ***\n\nSolution \nThe team realized that they did not need to build all of this\nfunctionality in-house and turned to [AutoML Vision Edge](/docs/ml/automl-image-labeling)\nto help them with this user experience challenge. \"There were\ntwo key considerations: how do we provide a seamless upload\nexperience for sellers to reduce friction for them when\ncreating listings, and how do we reduce the amount of cost,\nboth in terms of engineering effort and server costs, to\nsupport this type of feature,\" said Jake Hall, Head of Native\nApps for eBay Motors. AutoML addressed both issues head on,\n\"allowing us to improve the time it takes to create a listing\nthrough the app, which obviously has clear implications for\nus in terms of seller ROI,\" added Hall. \nThanks to AutoML Vision Edge, the eBay Motors team was able to\ngo from concept to prototype within a week using their own\ndata set. With just a few hours of data labeling, they were\nalso able to improve the model to a level that they feel\ncomfoable using in production.\n\n*** ** * ** ***\n\nResults\n\n1 week to go from concept to prototype (using their own data set)\n\n20 hours of data labeling needed to use the model in production \n\n\"AutoML allowed us to improve the time it takes to create a listing through the app, which obviously has clear implications for us in terms of seller ROI.\"\n\n\n- Jake Hall, Head of Native Apps, eBay Motors \nTry Firebase today\n\n\nIntegrating it into your app is easy.\n[Get started](https://console.firebase.google.com/) \n\nAll Firebase products \n\nBuild\n\n- [App Check](/products/app-check)\n- [App Hosting](/products/app-hosting)\n- [Authentication](/products/auth)\n- [Cloud Functions](/products/functions)\n- [Cloud Storage](/products/storage)\n- [Data Connect](/products/data-connect)\n- [Extensions](/products/extensions)\n- [Firestore](/products/firestore)\n- [Firebase ML](/products/ml)\n- [Genkit](https://genkit.dev/)\n- [Hosting](/products/hosting)\n- [Realtime Database](/products/realtime-database)\n- [Firebase AI Logic client SDKs](/products/firebase-ai-logic)\n\n[Generative AI](/products/generative-ai) \n\nRun\n\n- [A/B Testing](/products/ab-testing)\n- [App Distribution](/products/app-distribution)\n- [Cloud Messaging](/products/cloud-messaging)\n- [Crashlytics](/products/crashlytics)\n- [Google Analytics](/products/analytics)\n- [In-App Messaging](/products/in-app-messaging)\n- [Performance Monitoring](/products/performance)\n- [Remote Config](/products/remote-config)\n- [Test Lab](/products/test-lab)"]]
