Here you can find the code used for the training and experimental evaluation of the approach described in the paper "Edge-Aware Graph Matching Network forPart-based Semantic Segmentation" [1].

Abstract

Semantic segmentation of parts of objects is a marginally explored and challenging task in which multiple instances of objects and multiple parts within those objects must be recognized in an image. We introduce a novel approach (GMENet) for this task combining object-level context conditioning, part-level spatial relationships, and shape contour information. The first target is achieved by introducing a class-conditioning module that enforces class-level semantics when learning the part-level ones. In this way, mid-level features carry also this information prior to the decoding stage. To tackle part-level ambiguity and spatial relationships among parts we propose a novel adjacency graph-based module that aims at matching the spatial relationships between parts in the ground truth and predicted maps. Last, we introduce an additional module to further leverage edges localization. Besides testing our framework on the already used Pascal-Part 58 benchmark, we further introduce two novel benchmarks for large-scale part parsing, i.e., a more challenging version of Pascal-Part with 108 classes and the ADE20K-Part benchmark with 544 parts. GMENet achieves state-of-the-art results in all the considered tasks and furthermore allows to improve object-level segmentation accuracy.

 

Code and Dataset

The code for the training and the evaluation of the proposed method is available here

The datasets used for this work can be downloaded here.

Method

The overall architecture of the proposed approach is illustrated below.

 

Contacts

For any information on the method you can contact lttm@dei.unipd.it

Have a look at our website http://lttm.dei.unipd.it for other works on this topic.

 

 

References

[1]   U. Michieli and P. Zanuttigh, "Edge-Aware Graph Matching Network forPart-based Semantic Segmentation", under submission, 2022.

 

xhtml/css website layout by Ben Goldman  - http://realalibi.com