Deep Automatic Portrait Matting
نویسندگان
چکیده
We propose an automatic image matting method for portrait images. This method does not need user interaction, which was however essential in most previous approaches. In order to accomplish this goal, a new end-to-end convolutional neural network (CNN) based framework is proposed taking the input of a portrait image. It outputs the matte result. Our method considers not only image semantic prediction but also pixel-level image matte optimization. A new portrait image dataset is constructed with our labeled matting ground truth. Our automatic method achieves comparable results with state-of-the-art methods that require specified foreground and background regions or pixels. Many applications are enabled given the automatic nature of our system.
منابع مشابه
An Automatic Method for Image Matting Based on Saliency Detection ⋆
Image matting aims at extracting foreground from a given image by means of color and alpha estimation. An automatic matting method based on saliency detection was proposed in this paper, which takes the advantages of saliency detection as prior information instead of user interaction. After getting the saliency map, a cost function was constructed to obtain an alpha matte. Moreover, with the in...
متن کاملImproving alpha matte with depth information
In this paper, we propose an automatic alpha matting method that improves the quality of alpha matte using a depth camera. Depth camera obtains both visual and depth information of moving objects in real time. Our matting method improves existing natural alpha matting methods by adding depth information. Experimental results demonstrate that our matting method generates better results than the ...
متن کاملA novel framework for automatic trimap generation using the Gestalt laws of grouping
In this paper, we are concerned with unsupervised natural image matting. Due to the under-constrained nature of the problem, image matting algorithms are usually provided with user interactions, such as scribbles or trimaps. This is a very tedious task and may even become impractical for some applications. For unsupervised matte calculation, we can either adopt a technique that supports an unsu...
متن کاملVideo Matting from Depth Maps
Image matting is the process of taking an image, isolating the foreground, and replacing the background with a new image. This can be a hard problem when the background is unknown as it is fundamentally unconstrained. We look at an existing technique for foreground-background separation called Bayesian matting and improve upon it by adding depth information acquired by a time-of-flight range sc...
متن کاملTOM-Net: Learning Transparent Object Matting from a Single Image
This paper addresses the problem of transparent object matting. Existing image matting approaches for transparent objects often require tedious capturing procedures and long processing time, which limit their practical use. In this paper, we first formulate transparent object matting as a refractive flow estimation problem. We then propose a deep learning framework, called TOM-Net, for learning...
متن کامل