AI Green Screen

Background removal with the latest AI technology! Pick any image or video and detect objects and background automatically - and not only for background removal, but for various other cool effects too. The app is based on semantic image segmentation, which is the concept of finding objects and boundaries in images. Google Research DeepLab is a state-of-art deep learning neural network for the semantic image segmentation - and now with AI Green Screen this awesome technology is available as an easy app for everyday use. Simply let AI detect the image objects and pick the effect to apply. True power of artificial intelligence to everybody!

AI Detection Sensitivity

You can apply effects to detected objects or everything else (the background). There are 3 different levels of object detection sensitivity:

1) People only: strictly focusing on humans.

2) People and vehicles (default): detects people, airplanes, bicycles, boats, buses, cars, motorbikes, and trains. This mode is useful for protecting privacy, like blurring all objects that can contain privacy sensitive information (like license plates etc).

3) All objects: in addition to people and vehicles, detects: birds, bottles, cats, chairs, cows, diningtables, dogs, horses, pottedplants, sheeps, sofas, trains, and TVs

Effects

You can apply effects either to the detected objects or to the background: AI visualization, blurring, tiling, green screen (masking objects or background with green), transparency (background removal - transparency effects switch on image PNG saving), background swap, and many more.

Video support

Output frame rate control. Audio is not affected by the reduced frame rates. You can also trim the source video to process only a part of the footage.

Swap background video

It's fully automated green screening and replacing background with a selected video! This works for both images and videos. If you swap an image background with a video, you'll get an image-to-video export with audio taken from the background video.

Image to video effects

You can turn your image into short videos with looping and fluid animation effects. Just pick any photo and get super-cool seamless loops for Instagram, TikTok, etc!

Chroma keying

Use chroma key to process traditional green screen material - or build a mask based on image alpha (transparency).

Stickers

Save any project as a sticker for later use. Overlay stickers on top of images, image-to-video effects, or videos.

Create custom WhatsApp stickers

You can use all your custom stickers in WhatsApp too!

Use app as a share target

You can send media to AI Green Screen from any app that handles images or videos. Simply choose to share the selected image or video and use AI Green Screen as the share target.

Privacy

AI Green Screen uses local AI processing only, unlike most AI products. This is a major benefit, since AI Green Screen never sends your images out of your device for cloud processing. After editing, you can choose to share the image, if you like. But even in this case the standard Android share functionality is used, and you stay fully in control.

Mask Editor

Image mode: Sometimes AI may not produce a perfect result - or maybe you want to remove a person or an object from the AI generated mask. Extending the mask is just as easy with Mask Editor.

Crop Tool

Image mode: Choose to save only a selected part of the processed image.

Technology

Google TensorFlow Lite

https://www.tensorflow.org/lite/

Google Research DeepLab V3+ (PASCAL VOC 2012, Mobile Net V2)

https://github.com/tensorflow/models/tree/master/research/deeplab

We would love to hear your feedback. You can use Play to contact us. Or simply contact us on Twitter or Instagram @Haavepaja https://twitter.com/Haavepaja

使用最新的AI技术清除背景！挑选任何图像或视频并自动检测对象和背景-不仅用于去除背景，而且还用于其他各种炫酷效果。该应用程序基于语义图像分割，这是在图像中查找对象和边界的概念。 Google Research DeepLab是用于语义图像分割的最先进的深度学习神经网络-现在，借助AI Green Screen，该超赞的技术可以作为日常使用的便捷应用程序使用。只需让AI检测图像对象并选择要应用的效果即可。人工智能对每个人的真正力量！

AI检测灵敏度

您可以将效果应用于检测到的对象或其他所有东西（背景）。物体检测灵敏度分为3个级别：

1）只限人：严格关注人。

2）人员和车辆（默认）：检测人员，飞机，自行车，轮船，公共汽车，汽车，摩托车和火车。此模式对于保护隐私很有用，例如模糊可能包含隐私敏感信息的所有对象（例如车牌等）。

3）所有物体：除了人和车辆之外，还检测到：鸟，瓶子，猫，椅子，牛，餐桌，狗，马，盆栽，绵羊，沙发，火车和电视

效果

您可以将效果应用到检测到的对象或背景：AI可视化，模糊，平铺，绿屏（用绿色遮盖对象或背景），透明度（背景去除-透明效果打开图像PNG保存），背景交换和还有很多。

视频支持

输出帧率控制。降低的帧频不会影响音频。您还可以修剪源视频以仅处理部分镜头。

交换背景视频

这是全自动的绿色筛选，并用选定的视频替换背景！这适用于图像和视频。如果您将图像背景与视频交换，则会获得从背景视频获取的音频到图像到视频的导出。

图像到视频效果

您可以将图像转换成带有循环和流畅动画效果的短视频。只需选择任何照片，即可为Instagram，TikTok等获取超酷的无缝循环！

色度键控

使用色度键可处理传统的绿色屏幕材料-或基于图像alpha（透明度）构建蒙版。

贴纸

将任何项目另存为标签，以备后用。在图像，图像到视频效果或视频上覆盖贴纸。

创建自定义WhatsApp贴纸

您也可以在WhatsApp中使用所有自定义贴纸！

将应用用作共享目标

您可以从任何处理图像或视频的应用程序将媒体发送到AI绿屏。只需选择共享选定的图像或视频，然后使用AI绿屏作为共享目标。

隐私权

与大多数AI产品不同，AI Green Screen仅使用本地AI处理。这是一个主要好处，因为AI绿屏绝不会将图像发送出设备进行云处理。编辑后，您可以根据需要选择共享图像。但是即使在这种情况下，也使用了标准的Android共享功能，您可以完全控制自己。

遮罩编辑器

图像模式：有时AI可能无法产生完美的效果-或您可能想从AI生成的蒙版中移除人或物体。使用“蒙版编辑器”（Mask Editor），可以轻松扩展蒙版。

裁剪工具

图像模式：选择仅保存已处理图像的选定部分。

技术

Google TensorFlow Lite

https://www.tensorflow.org/lite/

Google Research DeepLab V3 +（PASCAL VOC 2012，Mobile Net V2）

https://github.com/tensorflow/models/tree/master/research/deeplab

我们很乐意倾听您的反馈。您可以使用Play与我们联系。或只是通过Twitter或Instagram @Haavepaja与我们联系https://twitter.com/Haavepaja