Learning Efficient Point Cloud Generation for Dense 3D Object Reconstruction

Authors: Chen-Hsuan Lin, Chen Kong, Simon Lucey

AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results for single-image 3D object reconstruction tasks show that we outperforms state-of-the-art methods in terms of shape similarity and prediction density. Our experimental results show that we generate much denser and more accurate shapes than state-of-the-art 3D prediction methods. We evaluate our proposed method by analyzing its performance in the application of single-image 3D reconstruction and comparing against state-of-the-art methods.
Researcher Affiliation Academia Chen-Hsuan Lin, Chen Kong, Simon Lucey The Robotics Institute Carnegie Mellon University chlin@cmu.edu, {chenk,slucey}@cs.cmu.edu
Pseudocode No The paper does not contain any pseudocode or algorithm blocks.
Open Source Code No The paper does not include any statement about releasing its source code, nor does it provide a link to a code repository.
Open Datasets Yes We train and evaluate all networks using the Shape Net database (Chang et al. 2015), which contains a large collection of categorized 3D CAD models.
Dataset Splits No The paper mentions '80%-20% training/test split' but does not specify a validation dataset split.
Hardware Specification No The paper vaguely mentions 'high-end GPUnodes' but does not provide specific hardware details (e.g., GPU models, CPU types, memory) used for running its experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers, such as libraries or frameworks used in implementation.
Experiment Setup Yes The structure generator predicts a 4N-channel image, which consists of the x, y, z coordinates and the binary mask from each of the N fixed viewpoint. We chose N = 8 with those viewpoints looking from the 8 corners of a centered cube. Orthographic projection is assumed in the transformation in (1) and (2). We take a two-stage training procedure: the structure generator is first pretrained to predict the x, y regular grids and depth images (z) from the N viewpoints (also pre-rendered with size 128 128), and then the network is fine-tuned with joint 2D projection optimization.