CNN Performance

Convolutional Neural Networks (CNNs) are a cornerstone of modern image analysis, but their performance is impacted by various factors, driving current research. Active areas of investigation include improving CNN robustness against adversarial attacks and mitigating performance drops due to data inconsistencies (e.g., scanner variations in medical imaging or non-uniform illumination). Researchers are exploring architectural improvements, such as incorporating attention mechanisms, physics-guided rules, and optimized activation functions, to enhance accuracy and efficiency, while also developing methods for model compression and training-free optimization. These advancements are crucial for deploying reliable and efficient CNNs in diverse applications, from medical diagnosis to autonomous systems.

Papers

October 2, 2024

Impact of White-Box Adversarial Attacks on Convolutional Neural Networks
Rakesh Podder, Sudipto Ghosh
Convolutional Neural Network Global Impact Adversarial Threat White Box Adversarial Attack CNN Performance White Box Adversarial

September 6, 2024

The Impact of Scanner Domain Shift on Deep Learning Performance in Medical Imaging: an Experimental Study
Brian Guo, Darui Lu, Gregory Szumel, Rongze Gui, Tingyu Wang, Nicholas Konz, Maciej A. Mazurowski
Deep Learning Deep Neural Network Global Impact Medical Image Domain Shift Medical Imaging Experimental Study CNN Performance

September 5, 2024

Non-Uniform Illumination Attack for Fooling Convolutional Neural Networks
Akshay Jain, Shiv Ram Dubey, Satish Kumar Singh, KC Santosh, Bidyut Baran Chaudhuri
Convolutional Neural Network CNN Performance ImageNet Attack Lightness Adaptation

September 3, 2024

Physical Rule-Guided Convolutional Neural Network
Kishor Datta Gupta, Marufa Kamal, Rakib Hossain Rifat, Mohd Ariful Haque, Roy George
Convolutional Neural Network Physic Informed CNN Performance M2m GNN

May 30, 2024

Optimizing cnn-Bigru performance: Mish activation and comparative analysis with Relu
Asmaa Benchama, Khalid Zebbara
Neural Network Deep Learning Comparative Study Intrusion Detection System ReLU Layer CNN Performance Capturing Extreme Activation

March 24, 2024

PaPr: Training-Free One-Step Patch Pruning with Lightweight ConvNets for Faster Inference
Tanvir Mahmud, Burhaneddin Yaman, Chun-Hao Liu, Diana Marculescu
High Performing 1D ConvNet Fast Inference Patch Level CNN Performance Lightweight ConvNets

March 15, 2024

When Training-Free NAS Meets Vision Transformer: A Neural Tangent Kernel Perspective
Qiqi Zhou, Yichen Zhu
Vision Transformer Neural Tangent Kernel Training Free CNN Performance

January 11, 2024

Self Expanding Convolutional Neural Networks
Blaise Appolinary, Alex Deaconu, Sophia Yang, Qingze, Li
Deep Learning Convolutional Neural Network Deep Convolutional Neural Network Self Information CNN Performance Sustainable Deep Learning

January 7, 2024

On The Potential of The Fractal Geometry and The CNNs Ability to Encode it
Julia El Zini, Bassel Musharrafieh, Mariette Awad
Deep Network Full Potential Fractal Structure CNN Performance Fractal Dimension Fractal Feature

November 13, 2023

PadChannel: Improving CNN Performance through Explicit Padding Encoding
Juho Kim
Convolutional Neural Network CNN Architecture Zero Padding CNN Performance

October 20, 2023

Using Human-like Mechanism to Weaken Effect of Pre-training Weight Bias in Face-Recognition Convolutional Neural Network
Haojiang Ying, Yi-Fan Li, Yiyang Chen
Convolutional Neural Network Mixed Effect CNN Model Human Like Balancing Weight CNN Performance

September 24, 2023

Evolution of Convolutional Neural Network (CNN): Compute vs Memory bandwidth for Edge AI
Dwith Chenna
Convolutional Neural Network Memory Trace CNN Network CNN Architecture CNN Performance Edge Artificial Intelligence

August 16, 2023

Exploring Winograd Convolution for Cost-effective Neural Network Fault Tolerance
Xinghua Xue, Cheng Liu, Bo Liu, Haitong Huang, Ying Wang, Tao Luo, Lei Zhang, Huawei Li, Xiaowei Li
CNN Performance Winograd Convolution Fault Aware Fault Tolerant Neural Network

August 2, 2023

Bio+Clinical BERT, BERT Base, and CNN Performance Comparison for Predicting Drug-Review Satisfaction
Yue Ling
Natural Language Processing Domain Specific CNN Performance Clinical BERT Drug Review

July 4, 2023

Toward more frugal models for functional cerebral networks automatic recognition with resting-state fMRI
Lukman Ismaila, Pejman Rasti, Jean-Michel Lemée, David Rousseau
Brain Network Visual Recognition Resting State Classical Convolutional Neural Network CNN Performance Five Dollar Model

April 30, 2023

Revealing Similar Semantics Inside CNNs: An Interpretable Concept-based Comparison of Feature Spaces
Georgii Mikriukov, Gesina Schwalbe, Christian Hellert, Korinna Bade
Convolutional Neural Network Comparative Study Feature Space CNN Architecture Interpretable Way CNN Performance Layer Similarity

October 14, 2022

Parameter-Free Average Attention Improves Convolutional Neural Network Performance (Almost) Free of Charge
Nils Körber
Attention Mechanism Computer Vision Task Visual Perception Curious Price Salient Region CNN Performance Parameter Free Attention

September 6, 2022

Improving the Accuracy and Robustness of CNNs Using a Deep CCA Neural Data Regularizer
Cassidy Pirlot, Richard C. Gerum, Cory Efird, Joel Zylberberg, Alona Fyshe
Neural Network Convolutional Neural Network Native Robustness Canonical Correlation Analysis Neural Network Representation CNN Performance

April 4, 2022

Feature robustness and sex differences in medical imaging: a case study in MRI-based Alzheimer's disease detection
Eike Petersen, Aasa Feragen, Maria Luise da Costa Zemsch, Anders Henriksen, Oskar Eiler Wiese Christensen, Melanie Ganz
Convolutional Neural Network Medical Imaging Disease Detection Magnetic Resonance CNN Performance Gender Difference ADNI Dataset

January 4, 2022

Attention Mechanism Meets with Hybrid Dense Network for Hyperspectral Image Classification
Muhammad Ahmad, Adil Mehmood Khan, Manuel Mazzara, Salvatore Distefano, Swalpa Kumar Roy, Xin Wu
Convolutional Neural Network Attention Mechanism Hyperspectral Image Classification Dense Network CNN Performance Parallel Attention Kernel Convolutional Neural Network

CNN Performance

Papers

Impact of White-Box Adversarial Attacks on Convolutional Neural Networks

The Impact of Scanner Domain Shift on Deep Learning Performance in Medical Imaging: an Experimental Study

Non-Uniform Illumination Attack for Fooling Convolutional Neural Networks

Physical Rule-Guided Convolutional Neural Network

Optimizing cnn-Bigru performance: Mish activation and comparative analysis with Relu

PaPr: Training-Free One-Step Patch Pruning with Lightweight ConvNets for Faster Inference

When Training-Free NAS Meets Vision Transformer: A Neural Tangent Kernel Perspective

Self Expanding Convolutional Neural Networks

On The Potential of The Fractal Geometry and The CNNs Ability to Encode it

PadChannel: Improving CNN Performance through Explicit Padding Encoding

Using Human-like Mechanism to Weaken Effect of Pre-training Weight Bias in Face-Recognition Convolutional Neural Network

Evolution of Convolutional Neural Network (CNN): Compute vs Memory bandwidth for Edge AI

Exploring Winograd Convolution for Cost-effective Neural Network Fault Tolerance

Bio+Clinical BERT, BERT Base, and CNN Performance Comparison for Predicting Drug-Review Satisfaction

Toward more frugal models for functional cerebral networks automatic recognition with resting-state fMRI

Revealing Similar Semantics Inside CNNs: An Interpretable Concept-based Comparison of Feature Spaces

Parameter-Free Average Attention Improves Convolutional Neural Network Performance (Almost) Free of Charge

Improving the Accuracy and Robustness of CNNs Using a Deep CCA Neural Data Regularizer

Feature robustness and sex differences in medical imaging: a case study in MRI-based Alzheimer's disease detection

Attention Mechanism Meets with Hybrid Dense Network for Hyperspectral Image Classification