A Comparative Study Of RESNET-Pretrained Models For Computer Vision

The science of computer vision has recently seen a revolution with deep learning models, and Residual Networks (ResNets) have become one of the most popular and efficient designs. ResNets have outperformed conventional convolutional neural networks in several image classification tasks, demonstrating outstanding results. An overview of ResNets and their variants, including Res2Net, COVID-ResNet, ResNeSt, SENet, and others, is presented in this review. We assess the performance of the ResNet design variants using benchmark datasets, highlight the major ideas and technological advancements underlying them, and their pros and cons. We also discuss some recent developments in ResNets research and applications, including object identification, picture segmentation, and transfer learning. This review would be beneficial to the research community in selecting the appropriate models based on the intended usage.

Paper

Full text

PDF

A Comparative Study Of RESNET-Pretrained Models For Computer Vision

Semantic Scholar · Computer Science · 2023

Abstract

The science of computer vision has recently seen a revolution with deep learning models, and Residual Networks (ResNets) have become one of the most popular and efficient designs. ResNets have outperformed conventional convolutional neural networks in several image classification tasks, demonstrating outstanding results. An overview of ResNets and their variants, including Res2Net, COVID-ResNet, ResNeSt, SENet, and others, is presented in this review. We assess the performance of the ResNet design variants using benchmark datasets, highlight the major ideas and technological advancements underlying them, and their pros and cons. We also discuss some recent developments in ResNets research and applications, including object identification, picture segmentation, and transfer learning. This review would be beneficial to the research community in selecting the appropriate models based on the intended usage.

Similar papers

© 2026 NYSGPT2525 LLC