ShuffleTrans: Patch-wise weight shuffle for transparent object segmentation

Transparent objects widely exist in the world. The task of transparent object segmentation is challenging as the object lacks its own texture. The cue of shape information therefore gets more critical. Most existing methods, however, rely on the mechanism of simple convolution, which is good at local cues and performs weakly on global cues like shape. To solve this problem, an operation named Patch-wise Weight Shuffle is proposed to bring in the global context cue by being combined with the dynamic convolution. A network ShuffleTrans that recognizes shape better is then designed based on this operation. Besides, fitter for this task, two auxiliary modules are presented in ShuffleTrans: a Boundary and Direction Refinement Module which collects two additional information, and a Channel Attention Enhancement Module that assists the above operation. Experiments on four texture-less object segmentation datasets and two normal datasets verify the effectiveness and generality of the method. Especially, the ShuffleTrans achieved 74.93% mIoU on the Trans10k v2 test set, which is more accurate than existing methods. & COPY; 2023 Elsevier Ltd. All rights reserved.

Automated summary

Transparent object segmentation is a challenging task due to the lack of texture. Shape information plays a critical role in this task. To address this issue, the researchers propose an operation called Patch-wise Weight Shuffle and design a network called ShuffleTrans that performs better in shape recognition. Experimental results on multiple datasets demonstrate the effectiveness of the method in transparent object segmentation.