Towards Enhancing Fine-grained Details for Image Matting

TL;DR

This paper introduces a novel deep image matting model that enhances fine-grained details like hairs and furs by using a parallel texture path and feature fusion, outperforming previous methods especially in preserving microscopic details.

Contribution

The paper proposes a downsampling-free texture path and a feature fusion unit to better recover microscopic details in image matting, addressing limitations of existing encoder-decoder models.

Findings

Outperforms previous state-of-the-art on Composition-1k dataset.

Effectively preserves tiny details such as hairs and furs.

Improves robustness to trimap quality through novel loss and trimap generation methods.

Abstract

In recent years, deep natural image matting has been rapidly evolved by extracting high-level contextual features into the model. However, most current methods still have difficulties with handling tiny details, like hairs or furs. In this paper, we argue that recovering these microscopic details relies on low-level but high-definition texture features. However, {these features are downsampled in a very early stage in current encoder-decoder-based models, resulting in the loss of microscopic details}. To address this issue, we design a deep image matting model {to enhance fine-grained details. Our model consists of} two parallel paths: a conventional encoder-decoder Semantic Path and an independent downsampling-free Textural Compensate Path (TCP). The TCP is proposed to extract fine-grained details such as lines and edges in the original image size, which greatly enhances the fineness of prediction. Meanwhile, to leverage the benefits of high-level context, we propose a feature fusion unit(FFU) to fuse multi-scale features from the semantic path and inject them into the TCP. In addition, we have observed that poorly annotated trimaps severely affect the performance of the model. Thus we further propose a novel term in loss function and a trimap generation method to improve our model's robustness to the trimaps. The experiments show that our method outperforms previous start-of-the-art methods on the Composition-1k dataset.