An Automated CNN Recommendation System for Image Classification Tasks

TL;DR

This paper introduces an automated system that efficiently recommends the most suitable CNN model for image classification tasks without requiring model training, based on precise evaluation of task complexity and model ability.

Contribution

It presents a novel, fast recommendation system that evaluates task complexity and model performance to suggest optimal CNN architectures without training.

Findings

Evaluation methods are accurate and reliable

System recommends optimal CNN models quickly

No training needed for recommendations

Abstract

Nowadays the CNN is widely used in practical applications for image classification task. However the design of the CNN model is very professional work and which is very difficult for ordinary users. Besides, even for experts of CNN, to select an optimal model for specific task may still need a lot of time (to train many different models). In order to solve this problem, we proposed an automated CNN recommendation system for image classification task. Our system is able to evaluate the complexity of the classification task and the classification ability of the CNN model precisely. By using the evaluation results, the system can recommend the optimal CNN model and which can match the task perfectly. The recommendation process of the system is very fast since we don't need any model training. The experiment results proved that the evaluation methods are very accurate and reliable.