Skip Connection Blocks

Inverted Residual Block

Introduced by Sandler et al. in MobileNetV2: Inverted Residuals and Linear Bottlenecks

An Inverted Residual Block, sometimes called an MBConv Block, is a type of residual block used for image models that uses an inverted structure for efficiency reasons. It was originally proposed for the MobileNetV2 CNN architecture. It has since been reused for several mobile-optimized CNNs.

A traditional Residual Block has a wide -> narrow -> wide structure with the number of channels. The input has a high number of channels, which are compressed with a 1x1 convolution. The number of channels is then increased again with a 1x1 convolution so input and output can be added.

In contrast, an Inverted Residual Block follows a narrow -> wide -> narrow approach, hence the inversion. We first widen with a 1x1 convolution, then use a 3x3 depthwise convolution (which greatly reduces the number of parameters), then we use a 1x1 convolution to reduce the number of channels so input and output can be added.

Source: MobileNetV2: Inverted Residuals and Linear Bottlenecks


Paper Code Results Date Stars


Task Papers Share
Image Classification 86 11.70%
Object Detection 49 6.67%
Quantization 45 6.12%
Classification 41 5.58%
Semantic Segmentation 29 3.95%
General Classification 18 2.45%
Management 10 1.36%
Multi-Task Learning 10 1.36%
Instance Segmentation 10 1.36%