Identification of utility images on a mobile device

2020 ◽  
Vol 2020 (8) ◽  
pp. 269-1-269-6
Author(s):  
Karthick Shankar ◽  
Qian Lin ◽  
Jan Allebach
Keyword(s):  
Neural Networks ◽  
Low Power ◽  
Mobile Phones ◽  
Mobile Device ◽  
Power Devices ◽  
Trade Off ◽  
Photo Album ◽  
The Media ◽  
Grocery Products ◽  
Looking Back

Mobile phones are used ubiquitously to capture all kinds of images – food, travel, friends, family, receipts, documents, grocery products and many more. Often times when looking back on photos to relive memories, we want to see images that actually represent experiences and not quick convenience photos that were taken for note-keeping and not deleted. Thus, we need to have a solution that presents only the relevant pictures without showing images of receipts, grocery products etc. – termed in general as utility images. This is in the context of a photobook which compiles and shows relevant images from the photo album of a mobile device. Further, all this has to be done on a mobile device since all the media resides there – introducing the need for our system to work on low power devices. In this paper, we present a work that can distinguish between utility and non-utility images. We also present a dataset of utility images and non-utility images with images for each category mentioned. Furthermore, we present a comparison between accuracies of popular pre-trained neural networks and show the trade-off between size and accuracy.

scholarly journals Sistem Informasi Penjualan Mobil Berbasis Web dengan Memanfaatkan Metode Visual Architecting ProcessTM (Studi Kasus : CV. Cahaya Mobilindo)

2019 ◽  
Vol 9 (2) ◽  
Author(s):  
Hendra Di Kesuma
Keyword(s):  
Information System ◽  
Mobile Devices ◽  
Programming Language ◽  
Mobile Phones ◽  
Mobile Device ◽  
Software Tools ◽  
Visual Methods ◽  
Trade Off ◽  
The Right

<p align="center"><strong><em>Abstract</em></strong></p><p><em>This research explores how architectural needs can be analysed and priorities determined to create the right architecture. Moreover, validation of the architecture can be done at the same time to ensure that the architects and stakeholders agree that the chosen architecture is the most appropriate.</em></p><p><em>The purpose of this study is to analyze and design an application mobile phones, using a visual methods architecting process based on Flashlite<sup>TM</sup>, to help consumers wishing to buy a car to research their intended purchase. Software tools can be designed using the Visual Architecting Process<sup>TM</sup> and implemented with the programming language PHP and Actionscript. This study concludes that clients who use mobile devices can successfully interact with a server that contains detailed information about a car and register their interests with the seller at the same time.</em></p><p><em> </em><strong><em>Keywords:</em></strong><em> </em><em>V</em><em>isual Architecting Process™ Methods, flashlite</em><em>, information system</em></p><p><em> </em><strong><em>Abstrak</em></strong></p><p><em>Metode Visual Architecting Process™ yang diusulkan oleh Bredemeyer Consulting mencakup teknik-teknik yang meliputi pemodelan arsitektur dan analisa untung-rugi (trade-off) yang digunakan dalam pembuatan arsitektur secara teknis. Metodologi ini mencakup bagaimana menganalisa kebutuhan-kebutuhan arsitektur dan menentukan prioritasnya untuk menciptakan arsitektur yang benar sekaligus melakukan validasi arsitektur sehingga memastikan bahwa arsitek dan stakeholders setuju bahwa arsitektur yang dihasilkan sungguh-sungguh arsitektur yang benar.</em></p><p><em>Tujuan dari penelitian ini adalah merancang </em><em>suatu aplikasi penjualan mobil dengan memanfaatkan metode Visual Architechting Process yang mampu menyediakan layanan informasi yang ditanam pada handphon teknologi flashlite dan m</em><em>emberikan kemudahan dalam hal berpromosi dan melakukan perhitungan kredit.</em></p><p><em>Dalam penelitian ini disimpulkan bahwa</em><em> client yang menggunakan mobile device dapat berinteraksi dengan server yang memuat informasi tentang mobil secara lengkap dan dapat melakukan registrasi secara langsung pada saat ingin  memesan dengan memanfaatkan service yang dibangun. Perangkat lunak dapat dibuat dengan menggunakan perancangan dengan metode Visual Architecting Process™ dan mengimplementasikannya dengan bahasa pemrograman PHP dan Actionscript.</em></p><strong><em>Kata kunci :</em></strong><em> Metode Visual Architecting Process™, flashlite, sistem informasi</em>

scholarly journals Quantization and Deployment of Deep Neural Networks on Microcontrollers

Sensors ◽  
2021 ◽  
Vol 21 (9) ◽  
pp. 2984
Author(s):  
Pierre-Emmanuel Novac ◽  
Ghouthi Boukli Hacene ◽  
Alain Pegatoquet ◽  
Benoît Miramond ◽  
Vincent Gripon
Keyword(s):  
Neural Networks ◽  
Low Power ◽  
Power Efficiency ◽  
Deep Neural Networks ◽  
Use Cases ◽  
Comparison Study ◽  
Power Devices ◽  
Embedded Devices ◽  
Inference Engines ◽  
New Framework

Embedding Artificial Intelligence onto low-power devices is a challenging task that has been partly overcome with recent advances in machine learning and hardware design. Presently, deep neural networks can be deployed on embedded targets to perform different tasks such as speech recognition, object detection or Human Activity Recognition. However, there is still room for optimization of deep neural networks onto embedded devices. These optimizations mainly address power consumption, memory and real-time constraints, but also an easier deployment at the edge. Moreover, there is still a need for a better understanding of what can be achieved for different use cases. This work focuses on quantization and deployment of deep neural networks onto low-power 32-bit microcontrollers. The quantization methods, relevant in the context of an embedded execution onto a microcontroller, are first outlined. Then, a new framework for end-to-end deep neural networks training, quantization and deployment is presented. This framework, called MicroAI, is designed as an alternative to existing inference engines (TensorFlow Lite for Microcontrollers and STM32Cube.AI). Our framework can indeed be easily adjusted and/or extended for specific use cases. Execution using single precision 32-bit floating-point as well as fixed-point on 8- and 16 bits integers are supported. The proposed quantization method is evaluated with three different datasets (UCI-HAR, Spoken MNIST and GTSRB). Finally, a comparison study between MicroAI and both existing embedded inference engines is provided in terms of memory and power efficiency. On-device evaluation is done using ARM Cortex-M4F-based microcontrollers (Ambiq Apollo3 and STM32L452RE).

Sky Segmentation for Enhanced Depth Reconstruction and Bokeh Rendering with Efficient Architectures

2020 ◽  
Vol 2020 (14) ◽  
pp. 378-1-378-7
Author(s):  
Tyler Nuanes ◽  
Matt Elsey ◽  
Radek Grzeszczuk ◽  
John Paul Shen
Keyword(s):  
Real Time ◽  
Mobile Device ◽  
Computational Cost ◽  
False Positives ◽  
Compact Model ◽  
High Quality ◽  
False Negatives ◽  
Trade Off ◽  
Depth Reconstruction ◽  
Binary Classifiers

We present a high-quality sky segmentation model for depth refinement and investigate residual architecture performance to inform optimally shrinking the network. We describe a model that runs in near real-time on mobile device, present a new, highquality dataset, and detail a unique weighing to trade off false positives and false negatives in binary classifiers. We show how the optimizations improve bokeh rendering by correcting stereo depth misprediction in sky regions. We detail techniques used to preserve edges, reject false positives, and ensure generalization to the diversity of sky scenes. Finally, we present a compact model and compare performance of four popular residual architectures (ShuffleNet, MobileNetV2, Resnet-101, and Resnet-34-like) at constant computational cost.

scholarly journals Neural Networks: Low‐Power Self‐Rectifying Memristive Artificial Neural Network for Near Internet‐of‐Things Sensor Computing (Adv. Electron. Mater. 6/2021)

2021 ◽  
Vol 7 (6) ◽  
pp. 2170017
Author(s):  
Seok Choi ◽  
Yong Kim ◽  
Tien Van Nguyen ◽  
Won Hee Jeong ◽  
Kyeong‐Sik Min ◽  
...  
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Artificial Neural Network ◽  
Internet Of Things ◽  
Low Power ◽  
Artificial Neural

scholarly journals DESIGN OF RF ENERGY HARVESTING SYSTEM FOR ENERGIZING LOW POWER DEVICES

2012 ◽  
Vol 132 ◽  
pp. 49-69 ◽  
Author(s):  
Norashidah Md. Din ◽  
Chandan Kumar Chakrabarty ◽  
Aima Bin Ismail ◽  
Kavuri Kasi Annapurna Devi ◽  
Wan-Yu Chen
Keyword(s):  
Energy Harvesting ◽  
Low Power ◽  
Power Devices ◽  
Rf Energy Harvesting ◽  
Energy Harvesting System ◽  
Rf Energy
Sign in / Sign up

Export Citation Format

Share Document