Going Deeper with Convolutions

TL;DR

The paper introduces the Inception architecture, notably GoogLeNet, a deep convolutional neural network that achieved state-of-the-art results in ImageNet 2014 by efficiently utilizing computational resources through multi-scale processing.

Contribution

It presents a novel deep CNN architecture that increases depth and width without extra computational cost, improving image classification and detection performance.

Findings

Achieved new state-of-the-art in ImageNet classification and detection

Designed a 22-layer deep network called GoogLeNet

Demonstrated efficient resource utilization in deep networks

Abstract

We propose a deep convolutional neural network architecture codenamed "Inception", which was responsible for setting the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC 2014). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. This was achieved by a carefully crafted design that allows for increasing the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC 2014 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.