Cut Cut is a professional image editing app. It is produced by APUS and integrates functions such as automatic AI image cutouts, easy background switchovers, and professional image editing. It also provides artistic camera filters and libraries with a wide range of materials. You can edit images using this app and post them to social media platforms such as WhatsApp and Facebook.
Image segmentation is one of the core features provided by Cut Cut. This feature allows users to cut out the portrait from an image and paste it onto any background. Before integrating the image segmentation service of HUAWEI Machine Learning (ML) Kit, Cut Cut was faced with the following multiple business challenges and was in urgent need of a high-performance, high-reliability, and high-quality algorithm solution.
High cost of calling on-cloud algorithms: When apps use on-cloud algorithms for image segmentation, they are calling third-party APIs. As the number of users increases, the calling cost and server cost of the algorithms also increase.
User concerns about privacy security during data transmission to the cloud: Images need to be uploaded to the cloud if on-cloud algorithms are used. Some users have concerns about the security of their private data when uploading their photos to the cloud.
App upgrade required for on-device algorithm upgrade: Apps can also integrate on-device algorithms for image segmentation. However, algorithms provided by third parties need to be fixed in apps. If the algorithms are upgraded, developers must release their apps again. As a result, users have to download the latest app again, compromising user experience.
With the release of HMS Core 4.0, HUAWEI ML Kit has launched its image segmentation service. It provides developers with the capability of identifying pixels of each kind of element in images or video frames. It also allows apps to implement functions such as image area optimization, portrait coloring, and sky filter effects based on the image segmentation result. The image segmentation service of HUAWEI ML Kit perfectly matches the business requirements of Cut Cut.
Free service and no limit on the number of calls: HMS Core is dedicated to building a new smart ecosystem covering all scenarios. On-device image segmentation algorithms provided by HUAWEI ML Kit are free and do not place a limit on the number of calls. HUAWEI ML Kit will obtain continuous investment due to its status as a core and strategic business so that it can help more developers to achieve business innovation and success.
On-device image processing without the risks of private data being leaked: The image segmentation service provided by HUAWEI ML Kit supports on-device image processing and is not affected by the users' network environment. It allows algorithms to run smoothly and accurately at any time anywhere on user devices without any risks of private data being leaked caused by the uploading of data.
One-time integration, algorithm upgrade with HMS Core upgrade: With one-time integration with HUAWEI ML Kit image segmentation, developers get the benefits of continuous algorithm model upgrades. Because the algorithm model in HUAWEI ML Kit is dynamically upgraded with HMS Core, developers do not need to release apps again if they want to upgrade the algorithm model.

To Cut Cut:
The image segmentation service of HUAWEI ML Kit is free and supports dynamic upgrades. This reduces maintenance and service costs required by Cut Cut. In addition, the service enables Cut Cut to quickly and accurately segment images. It also supports image segmentation for 11 image types, such as images of sky and grass, and supports video segmentation, greatly improving its competitiveness and user experience.
To users:
Precise image segmentation is available for users even when their devices do not have an Internet connection. This ensures a faster and better image cutout experience.
In addition to image segmentation, ML Kit also provides services such as face detection, text recognition, object detection and tracking, image classification, and landmark recognition. For more information and the demo app, visit ML Kit.