Sparse Convolutional

Sparse convolutional networks aim to improve the efficiency and performance of convolutional neural networks (CNNs) by leveraging the inherent sparsity present in many datasets, such as point clouds or video frames. Current research focuses on optimizing sparse convolution operations through techniques like binarization, adaptive masking, and the development of specialized hardware architectures (e.g., "Phantom") to accelerate inference. These advancements are significant because they enable the application of deep learning to large-scale, high-dimensional data, previously intractable due to computational limitations, with applications ranging from 3D scene reconstruction and autonomous driving to medical image analysis and neutrino telescope data processing.

Papers

May 4, 2023

Smaller3d: Smaller Models for 3D Semantic Segmentation Using Minkowski Engine and Knowledge Distillation Methods
Alen Adamyan, Erik Harutyunyan
Knowledge Distillation 3D Semantic Segmentation Small Model 3D Convolution 3D Deep Learning Sparse Convolutional

March 27, 2023

Binarizing Sparse Convolutional Networks for Efficient Point Cloud Analysis
Xiuwei Xu, Ziwei Wang, Jie Zhou, Jiwen Lu
Neural Network Sparse Convolution Sparse Convolutional

March 25, 2023

Adaptive Sparse Convolutional Networks with Global Context Enhancement for Faster Object Detection on Drone Images
Bowei Du, Yecheng Huang, Jiaxin Chen, Di Huang
Foreground Object Sparse Convolution Sparse Convolutional Drone Image

March 15, 2023

Trigger-Level Event Reconstruction for Neutrino Telescopes Using Sparse Submanifold Convolutional Neural Networks
Felix J. Yu, Jeffrey Lazar, Carlos A. Argüelles
Convolutional Neural Network Event Based Video Sparse Convolutional Trigger Free Event Detection Neutrino Telescope

December 17, 2022

FSCNN: A Fast Sparse Convolution Neural Network Inference System
Bo Ji, Tianyi Chen
Neural Network CNN Architecture Structured Sparsity Fast Convolution Sparse Convolutional

December 6, 2022

Automated Segmentation of Computed Tomography Images with Submanifold Sparse Convolutional Networks
Saúl Alonso-Monsalve, Leigh H. Whitehead, Adam Aurisano, Lorena Escudero Sanchez
CT Image Automatic Segmentation Random Sparsification Cancer Detection Sparse Convolutional Downsampling Method

October 24, 2022

Revisiting Sparse Convolutional Model for Visual Recognition
Xili Dai, Mingyang Li, Pengyuan Zhai, Shengbang Tong, Xingjian Gao, Shao-Lun Huang, Zhihui Zhu, Chong You, Yi Ma
Visual Recognition Sparse Regularization Sparse Convolutional Convolutional Sparse

April 26, 2022

Focal Sparse Convolutional Networks for 3D Object Detection
Yukang Chen, Yanwei Li, Xiangyu Zhang, Jian Sun, Jiaya Jia
3D Object Detection Sparse Convolution Focal Information Sparse Convolutional Sparse CNN

March 15, 2022

Interspace Pruning: Using Adaptive Filter Representations to Improve Training of Sparse CNNs
Paul Wimmer, Jens Mehnert, Alexandru Paul Condurache
Convolutional Neural Network Training Data Pruning Method Adaptive Filter Unstructured Pruning Sparse Convolutional Space Pruning

March 8, 2022

DeltaCNN: End-to-End CNN Inference of Sparse Frame Differences in Videos
Mathias Parger, Chengcheng Tang, Christopher D. Twigg, Cem Keskin, Robert Wang, Markus Steinberger
Convolutional Neural Network Gameplay Video CNN Inference Sparse Convolutional

February 21, 2022

PCSCNet: Fast 3D Semantic Segmentation of LiDAR Point Cloud for Autonomous Car using Point Convolution and Sparse Convolution Network
Jaehyun Park, Chansoo Kim, Kichun Jo
Semantic Segmentation Autonomous Vehicle 3D Semantic Segmentation Light Detection and Ranging Point Convolution Voxel Based Semantic Segmentation Sparse Convolutional

November 25, 2021

CIRCLE: Convolutional Implicit Reconstruction and Completion for Large-scale Indoor Scene
Haoxiang Chen, Jiahui Huang, Tai-Jiang Mu, Shi-Min Hu
Fine Grained 3D Scene Scene Completion 3D Room Sparse Convolutional Shape Refinement Green CIRCLe Reconstruction Method

November 9, 2021

Phantom: A High-Performance Computational Core for Sparse Convolutional Neural Networks
Mahmood Azhar Qureshi, Arslan Munir
Phantom 2D Accelerator Phantom Threat Sparse Convolutional Sparse Accelerator Sparse CNN

July 10, 2021

Dense-Sparse Deep Convolutional Neural Networks Training for Image Denoising
Basit O. Alawode, Mudassir Masood
Image Denoising Sparse Convolutional

Sparse Convolutional

Papers

Smaller3d: Smaller Models for 3D Semantic Segmentation Using Minkowski Engine and Knowledge Distillation Methods

Binarizing Sparse Convolutional Networks for Efficient Point Cloud Analysis

Adaptive Sparse Convolutional Networks with Global Context Enhancement for Faster Object Detection on Drone Images

Trigger-Level Event Reconstruction for Neutrino Telescopes Using Sparse Submanifold Convolutional Neural Networks

FSCNN: A Fast Sparse Convolution Neural Network Inference System

Automated Segmentation of Computed Tomography Images with Submanifold Sparse Convolutional Networks

Revisiting Sparse Convolutional Model for Visual Recognition

Focal Sparse Convolutional Networks for 3D Object Detection

Interspace Pruning: Using Adaptive Filter Representations to Improve Training of Sparse CNNs

DeltaCNN: End-to-End CNN Inference of Sparse Frame Differences in Videos

PCSCNet: Fast 3D Semantic Segmentation of LiDAR Point Cloud for Autonomous Car using Point Convolution and Sparse Convolution Network

CIRCLE: Convolutional Implicit Reconstruction and Completion for Large-scale Indoor Scene

Phantom: A High-Performance Computational Core for Sparse Convolutional Neural Networks

Dense-Sparse Deep Convolutional Neural Networks Training for Image Denoising