MobileNetV2: Inverted Residuals and Linear Bottlenecks

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.

BiSeNet: Bilateral Segmentation Network for Real-Time Semantic Segmentation

“…Following the pioneer work (Sandler et al, 2018;Howard et al, 2019), we design a Gather-and-Expansion Layer, as discussed in Section 4.2 and illustrated in Figure 5. The main improvements consist of two-fold: (i) we adopt one 3 × 3 convolution as the Gather Layer instead of one point-wise convolution in the inverted bottleneck of Mo-bileNetV2 (Sandler et al, 2018); (ii) when stride = 2, we employs two 3 × 3 depth-wise convolution to substitute a 5 × 5 depth-wise convolution. Table 4b shows the improvement of our block design.…”

“…It has an advantage in memory access cost (Sandler et al, 2018;Howard et al, 2019). The expansion ratio of can control the output dimension of this layer.…”

Binary neural networks: A survey

“…They visualize the feature maps extracted by different filters and view each filter as a visual unit focusing on different visual components.of the ResNet-50 [28], and meanwhile save more than 75% of parameters and 50% computational time. In the literature, approaches for compressing the deep networks can be classified into five categories: parameter pruning [26,29,30,31], parameter quantizing [32,33,34,35,36,37,38,39,40,41], low-rank parameter factorization [42,43,44,45,46], transferred/compact convolutional filters [47,48,49,50], and knowledge distillation [51,52,53,54,55,56]. The parameter pruning and quantizing mainly focus on eliminating the redundancy in the model parameters respectively by removing the redundant/uncritical ones or compressing the parameter space (e.g.…”

Block Annotation: Better Image Annotation With Sub-Image Decomposition

“…All existing results show performance with a VGG-16 based model. We train a MobileNet based model which has been shown to achieve similar performance to VGG-16 (71.8% vs 71.5% Top-1 accuracy on ImageNet) while requiring fewer computational resources [25,51]. Our fully-supervised implementation pretrained on ImageNet achieves 69.6% mIOU on Pascal VOC 2012 [17]; in comparison, the reference DeepLab-VGG16 model achieves 68.7% mIOU [12] and the re-implementation in [36] Performance Comparison.…”