Open AccessFilter Distribution Templates in Convolutional Networks for Image Classification Tasks
Author(s) -
Ramon Izquierdo-Cordova,
Walterio Mayol-Cuevas
Publication year - 2021
Language(s) - English
Resource type - Conference proceedings
DOI - 10.52591/lxai202106253
Subject(s) - computer science , convolutional neural network , residual neural network , filter (signal processing) , layer (electronics) , artificial intelligence , artificial neural network , pattern recognition (psychology) , network architecture , template , distribution (mathematics) , computer vision , mathematics , mathematical analysis , chemistry , computer security , organic chemistry , programming language
Neural network designers have reached progressive accuracy by increasing models depth, introducing new layer types and discovering new combinations of layers. A common element in many architectures is the distribution of the number of filters in each layer. Neural network models keep a pattern design of increasing filters in deeper layers such as those in LeNet, VGG, ResNet, MobileNet and even in automatic discovered architectures such as NASNet. It remains unknown if this pyramidal distribution of filters is the best for different tasks and constrains. In this work we present a series of modifications in the distribution of filters in three popular neural network models and their effects in accuracy and resource consumption. Results show that by applying this approach, some models improve up to 8.9% in accuracy showing reductions in parameters up to 54%.