MobileNetV2: Inverted Residuals and Linear Bottlenecks

TLDR

The paper introduces MobileNetV2, a mobile architecture that improves state‑of‑the‑art performance across tasks and model sizes, and presents efficient applications to object detection via SSDLite and to semantic segmentation via Mobile DeepLabv3. MobileNetV2 uses an inverted residual structure with shortcut connections between thin bottleneck layers, lightweight depthwise convolutions for non‑linearity, and decouples input/output domains from transformation expressiveness; the authors evaluate it on ImageNet, COCO, VOC, measuring accuracy, multiply‑add operations, latency, and parameters. MobileNetV2 achieves state‑of‑the‑art performance across tasks and model sizes, and the authors show that removing non‑linearities in narrow layers preserves representational power, improving performance.

Abstract

In this paper we describe a new mobile architecture, MobileNetV2, that improves the state of the art performance of mobile models on multiple tasks and benchmarks as well as across a spectrum of different model sizes. We also describe efficient ways of applying these mobile models to object detection in a novel framework we call SSDLite. Additionally, we demonstrate how to build mobile semantic segmentation models through a reduced form of DeepLabv3 which we call Mobile DeepLabv3. is based on an inverted residual structure where the shortcut connections are between the thin bottleneck layers. The intermediate expansion layer uses lightweight depthwise convolutions to filter features as a source of non-linearity. Additionally, we find that it is important to remove non-linearities in the narrow layers in order to maintain representational power. We demonstrate that this improves performance and provide an intuition that led to this design. Finally, our approach allows decoupling of the input/output domains from the expressiveness of the transformation, which provides a convenient framework for further analysis. We measure our performance on ImageNet [1] classification, COCO object detection [2], VOC image segmentation [3]. We evaluate the trade-offs between accuracy, and number of operations measured by multiply-adds (MAdd), as well as actual latency, and the number of parameters.