Learning global features by aggregating information over multiple views has
been shown to be effective for 3D shape analysis. For view aggregation in deep
learning models, pooling has been applied extensively. However, pooling leads
to a loss of the content within views, and the spatial relationship among
views, which limits the discriminability of learned features. We propose
3DViewGraph to resolve this issue, which learns 3D global features by more
effectively aggregating unordered views with attention. Specifically, unordered
views taken around a shape are regarded as view nodes on a view graph.
3DViewGraph first learns a novel latent semantic mapping to project low-level
view features into meaningful latent semantic embeddings in a lower dimensional
space, which is spanned by latent semantic patterns. Then, the content and
spatial information of each pair of view nodes are encoded by a novel spatial
pattern correlation, where the correlation is computed among latent semantic
patterns. Finally, all spatial pattern correlations are integrated with
attention weights learned by a novel attention mechanism. This further
increases the discriminability of learned features by highlighting the
unordered view nodes with distinctive characteristics and depressing the ones
with appearance ambiguity. We show that 3DViewGraph outperforms
state-of-the-art methods under three large-scale benchmarks.

本文提出了一种名为 3DViewGraph 的方法，该方法通过引入注意力机制，学习 3D 全局特征，解决了池化操作可能导致信息丢失和影响学习特征判别性的限制问题，并在三个大规模基准测试中展现出优异性能。