Searching for MobileNetV3

Howard, Andrew; Pang, Ruoming; Adam, Hartwig; Le, Quoc V.; Sandler, Mark; Chen, Bin; Wang, Weijun; Chen, Liang-Chieh; Tan, Mingxing; Chu, Grace; Vasudevan, Vijay; Zhu, Yukun

doi:10.1109/iccv.2019.00140

Cited by 6,201 publications

(3,183 citation statements)

References 31 publications

Supporting

Mentioning

2,545

Contrasting

Unclassified

Order By: Relevance

“…Block design of of Semantic Branch. 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.…”

Section: Ablative Evaluation On Cityscapesmentioning

confidence: 99%

See 1 more Smart Citation

BiSeNet: Bilateral Segmentation Network for Real-Time Semantic Segmentation

Wang

Peng³

et al. 2018

Lecture Notes in Computer Science

1,960

1,263

View full text Add to dashboard Cite

The low-level details and high-level semantics are both essential to the semantic segmentation task. However, to speed up the model inference, current approaches almost always sacrifice the low-level details, which leads to a considerable accuracy decrease. We propose to treat these spatial details and categorical semantics separately to achieve high accuracy and high efficiency for real-time semantic segmentation. To this end, we propose an efficient and effective architecture with a good trade-off between speed and accuracy, termed Bilateral Segmentation Network (BiSeNet V2). This architecture involves: (i) a Detail Branch, with wide channels and shallow layers to capture low-level details and generate high-resolution feature representation; (ii) a Semantic Branch, with narrow channels and deep layers to obtain high-level semantic context. The Semantic Branch is lightweight due to reducing the channel capacity and a fast-downsampling strategy. Furthermore, we design a Guided Aggregation Layer to enhance mutual connections and fuse both types of feature representation. Besides, a booster training strategy is designed to improve the segmentation performance without any extra inference cost. Extensive quantitative and qualitative evaluations demonstrate that the pro-posed architecture performs favourably against a few state-of-the-art real-time semantic segmentation approaches. Specifically, for a 2,048×1,024 input, we achieve 72.6% Mean IoU on the Cityscapes test set with a speed of 156 FPS on one NVIDIA GeForce GTX 1080 Ti card, which is significantly faster than existing methods, yet we achieve better segmentation accuracy. Code and trained models will be made publicly available.

show abstract

Section: Ablative Evaluation On Cityscapesmentioning

confidence: 99%

“…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.…”

Section: Ablative Evaluation On Cityscapesmentioning

confidence: 99%

BiSeNet: Bilateral Segmentation Network for Real-Time Semantic Segmentation

Wang

Peng³

et al. 2018

Lecture Notes in Computer Science

1,960

1,263

View full text Add to dashboard Cite

show abstract

“…While our current implementation focusses on large-scale image data from HT phenotyping platforms installed in controlled greenhouse environments, we envision the approach to also be beneficial for field phenotyping purposes. For example, choosing a smaller network backbone such as MobileNetV3 [31] could enable field researchers to measure plant morphometry on the go using their smartphone and could thus facilitate data collection and interpretation in the field.…”

Section: Outlook and Perspectivementioning

confidence: 99%

aradeepopsis: From images to phenotypic traits using deep transfer learning

Huether

Schandry

Jandrasits

et al. 2020

Preprint

View full text Add to dashboard Cite

Linking plant phenotype to genotype, i.e., identifying genetic determinants of phenotypic traits, is a common goal of both plant breeders and geneticists. While the ever-growing genomic resources and rapid decrease of sequencing costs have led to enormous amounts of genomic data, collecting phenotypic data for large numbers of plants remains a bottleneck. Many phenotyping strategies rely on imaging plants, which makes it necessary to extract phenotypic measurements from these images rapidly and robustly. Common image segmentation tools for plant phenotyping mostly rely on color information, which is error-prone when either background or plant color deviate from the underlying expectations. We have developed a versatile, fully open-source pipeline to extract phenotypic measurements from plant images in an unsupervised manner. ᴀʀᴀᴅᴇᴇᴘᴏᴘsɪs was built around the deep-learning model DeepLabV3+ that was re-trained for segmentation of Arabidopsis thaliana rosettes. It uses semantic segmentation to classify leaf tissue into up to three categories: healthy, anthocyaninrich, and senescent. This makes ᴀʀᴀᴅᴇᴇᴘᴏᴘsɪs particularly powerful at quantitative phenotyping from early to late developmental stages, of mutants with aberrant leaf color and/or phenotype, and of plants growing in stressful conditions where leaf color may deviate from green. Using our tool on a panel of 210 natural Arabidopsis accessions, we were able to not only accurately segment images of phenotypically diverse genotypes but also to map known loci related to anthocyanin production and early necrosis using the ᴀʀᴀᴅᴇᴇᴘᴏᴘsɪs output in genome-wide association analyses. Our pipeline is able to handle images of diverse origins, image quality, and background composition, and could even accurately segment images of a distantly related Brassicaceae. Because it can be deployed on virtually any common operating system and is compatible with several high-performance computing environments, ᴀʀᴀᴅᴇᴇᴘᴏᴘsɪs can be used independently of bioinformatics expertise and computing resources. ᴀʀᴀᴅᴇᴇᴘᴏᴘsɪs is available at https://github.com/Gregor-Mendel-Institute/aradeepopsis.

show abstract

“…Though the external servers are able to run deeper models than phones, it requires fast and stable internet connection at all times. Moreover, the advancement of efficient neural networks such as MobileNets enable a deep-learning approach to be implemented on a mobile device [22,13].…”

Section: Related Workmentioning

confidence: 99%

“…Specification for our neural network. The bneck operator denotes the bottleneck block as defined in [13]. k denotes the kernel size.…”

Section: Training Proceduresmentioning

confidence: 99%

Street Crossing Aid Using Light-Weight CNNs for the Visually Impaired

Lee

Kim

2019

2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW)

View full text Add to dashboard Cite

In this paper, we address an issue that the visually impaired commonly face while crossing intersections and propose a solution that takes form as a mobile application. The application utilizes a deep learning convolutional neural network model, LytNetV2, to output necessary information that the visually impaired may lack when without human companions or guide-dogs. A prototype of the application runs on iOS devices of versions 11 or above. It is designed for comprehensiveness, concision, accuracy, and computational efficiency through delivering the two most important pieces of information, pedestrian traffic light color and direction, required to cross the road in real-time. Furthermore, it is specifically aimed to support those facing financial burden as the solution takes the form of a free mobile application. Through the modification and utilization of key principles in MobileNetV3 such as depthwise seperable convolutions and squeeze-excite layers, the deep neural network model achieves a classification accuracy of 96% and average angle error of 6.15 • , while running at a frame rate of 16.34 frames per second. Additionally, the model is trained as an image classifier, allowing for a faster and more accurate model. The network is able to outperform other methods such as object detection and non-deep learning algorithms in both accuracy and thoroughness. The information is delivered through both auditory signals and vibrations, and it has been tested on seven visually impaired and has received above satisfactory responses.

show abstract

Searching for MobileNetV3

Cited by 6,201 publications

References 31 publications

BiSeNet: Bilateral Segmentation Network for Real-Time Semantic Segmentation

BiSeNet: Bilateral Segmentation Network for Real-Time Semantic Segmentation

aradeepopsis: From images to phenotypic traits using deep transfer learning

Street Crossing Aid Using Light-Weight CNNs for the Visually Impaired

Contact Info

Product

Resources

About