Rethinking the Inception Architecture for Computer Vision

TL;DR

This paper proposes a rethought Inception architecture for computer vision that improves efficiency and accuracy by using factorized convolutions and regularization, achieving state-of-the-art results with fewer parameters.

Contribution

The paper introduces a novel Inception-based network design that enhances computational efficiency and accuracy, setting new benchmarks on the ImageNet classification challenge.

Findings

Achieved 21.2% top-1 error with 5 billion multiply-adds.

Reduced model size to less than 25 million parameters.

Ensemble of models reached 3.5% top-5 error on validation.

Abstract

Convolutional networks are at the core of most state-of-the-art computer vision solutions for a wide variety of tasks. Since 2014 very deep convolutional networks started to become mainstream, yielding substantial gains in various benchmarks. Although increased model size and computational cost tend to translate to immediate quality gains for most tasks (as long as enough labeled data is provided for training), computational efficiency and low parameter count are still enabling factors for various use cases such as mobile vision and big-data scenarios. Here we explore ways to scale up networks in ways that aim at utilizing the added computation as efficiently as possible by suitably factorized convolutions and aggressive regularization. We benchmark our methods on the ILSVRC 2012 classification challenge validation set demonstrate substantial gains over the state of the art: 21.2% top-1 and 5.6% top-5 error for single frame evaluation using a network with a computational cost of 5 billion multiply-adds per inference and with using less than 25 million parameters. With an ensemble of 4 models and multi-crop evaluation, we report 3.5% top-5 error on the validation set (3.6% error on the test set) and 17.3% top-1 error on the validation set.