DeepPano: Deep Panoramic Representation for 3-D Shape Recognition

By Baoguang Shi, Song Bai, Zhichao Zhou, Xiang Bai. 2015

403 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, 3D Vision & Pose

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In 3D Vision & Pose

Citations in the corpus

Learning Descriptor Networks for 3D Shape Synthesis and Analysis (2018, neutral): “[23] 88.4% PANORAMA-NN [19] 91.1% ECC [22] 90.0% 3D ShapeNets [29] 83.5% DeepPano [21] 85.5% SPH [10] 79.8% VConv-DAE [20] 80.5% 3D-GAN [28] 91.0% 3D DescriptorNet (ours) 92.4 %”
Set Transformer (2018, methodology): “Pooling architectures for permutation invariant mappings Pooling architectures for sets have been used in various problems such as 3D shape recognition (Shi et al., 2015; Su et al., 2015), discovering causality (Lopez-Paz et al.”
Volumetric and Multi-view CNNs for Object Classification on 3D Data (2016, background): “In this context, DeepPano [25] converts 3D shapes into panoramic views; i.”
Multi-View Silhouette and Depth Decomposition for High Resolution 3D Object Representation (2018, neutral): “Multi-View Representation Our work connects to multi-view representations which capture the characteristics of a 3D object from multiple viewpoints in 2D [17, 26, 42, 32, 12, 39, 34], such as decomposing image silhouettes [23? ], Light Field Descriptors [2], and 2D panoramic mapping [38].”
Deep Sets (2017, background): “Interestingly, variations of view/angle-pooling ( e.g. Su et al., 2015; Shi et al., 2015) can be interpreted as set-pooling where the class-label is invariant to permutation of different views.”
FPNN: Field Probing Neural Networks for 3D Data (2016, methodology): “Since CNNs for images have been extensively studied, 3D shapes can be rendered into 2D images, and be represented by the CNN features of the images [25, 28], which, surprisingly, outperforms any 3D CNN approaches, in a 3D shape classification task.”
Orientation-boosted Voxel Nets for 3D Object Recognition (2016, neutral): “d. mean std. Recursive D [23] - - - - 37.6 - - - Recursive D+C [23] - - - - 44.8 - - - Triangle+SVM [17] 67.1 - - - - - - - GFH+SVM [24] 71.0 - - - - - - - ShapeNet [1] - - 83.5 - 57.9 - - - DeepPano [25] - - 85.5 - - - - - VoxNet [2] 72. - 92 - 71. - - - Our Baseline 74.8 0.0 92.8 0.3 75.5 0.1 93.0”
FusionNet: 3D Object Classification Using Multiple Data Representations (2016, methodology): “In [20], the authors suggest a new robust representation of 3D data by way of a cylindrical panoramic projection that is learned using CNN.”
Recurrent Slice Networks for 3D Segmentation of Point Clouds (2018, background): “In [29], 3D shapes were transformed into panoramic views, i.”
3D Object Classification via Spherical Projections (2017, methodology): “This is relevant to some recent works of classifying panoramas of 3D objects [22, 21].”
Learning a Hierarchical Latent-Variable Model of 3D Shapes (2017, methodology): “vised state-of-the-art, outperforming models such as 3D ShapeNet [45], DeepPano [38], and the geometry imagebase approach [39], by a large margin, and comes close to models such as VoxNet [27].”
LPD-Net: 3D Point Cloud Learning for Large-Scale Place Recognition and Environment Analysis (2018, methodology): “Some other methods, such as the DeepPano [18] and Multiview CNNs [20], project 3D point clouds into 2D images and use the 2D CNN to learn features.”
Pairwise Decomposition of Image Sequences for Active Multi-view Recognition (2016, background): “Recently, CNN architectures have been extended to allow for recognition from image sequences using a single network, by max pooling across all viewpoints [35], or by unwrapping an object shape into a panorama and max pooling across each row [33].”
Pointwise Convolutional Neural Networks (2017, background): “Recently, several approaches to using 3D convolutional networks to extract shape descriptor have been proposed, ranging from voxel-based representation [42, 20] panorama [32], feature pooling from 2D projections from multiple viewpoints [37, 27], to point set [27].”
Point Transformer (2020, positive): “Then, working with traditional CNNs is possible [12], [25].”