Detecting and reading text from natural images is a hard computer vision task that is central to a variety of emerging applications. Related problems like document character recognition have been widely studied by computer vision and machine learning researchers and are virtually solved for practical applications like reading handwritten digits. Reliably recognizing characters in more complex scenes like photographs, however, is far more difficult: the best existing methods lag well behind human performance on the same tasks. In this paper we attack the problem of recognizing digits in a real application using unsupervised feature learning methods: reading house numbers from street level photos. To this end, we introduce a new benchmark dataset for research use containing over 600,000 labeled digits cropped from Street View images. We then demonstrate the difficulty of recognizing these digits when the problem is approached with hand-designed features. Finally, we employ variants of two recently proposed unsupervised feature learning methods and find that they are convincingly superior on our benchmarks.

/pdf/reading-digits-in-natural-images-with-unsupervised-feature-3ckuq1665o.pdf

Reading Digits in Natural Images with Unsupervised Feature Learning

From the Publisher: 
The accessible presentation of this book gives both a general view of the entire computer vision enterprise and also offers sufficient detail to be able to build useful applications. Users learn techniques that have proven to be useful by first-hand experience and a wide range of mathematical methods. A CD-ROM with every copy of the text contains source code for programming practice, color images, and illustrative movies. Comprehensive and up-to-date, this book includes essential topics that either reflect practical significance or are of theoretical importance. Topics are discussed in substantial and increasing depth. Application surveys describe numerous important application areas such as image based rendering and digital libraries. Many important algorithms broken down and illustrated in pseudo code. Appropriate for use by engineers as a comprehensive reference to the computer vision enterprise.

Computer vision : a modern approach = 计算机视觉 : 一种现代的方法

IEEE Transactions on Pattern Analysis and Machine Intelligence

/pdf/a-break-in-the-clouds-towards-a-cloud-definition-svovsw1r5a.pdf

A Break in the Clouds: Towards a Cloud Definition

Robust detection and tracking of objects is crucial for the deployment of autonomous vehicle technology. Image based benchmark datasets have driven development in computer vision tasks such as object detection, tracking and segmentation of agents in the environment. Most autonomous vehicles, however, carry a combination of cameras and range sensors such as lidar and radar. As machine learning based methods for detection and tracking become more prevalent, there is a need to train and evaluate such methods on datasets containing range sensor data along with images. In this work we present nuTonomy scenes (nuScenes), the first dataset to carry the full autonomous vehicle sensor suite: 6 cameras, 5 radars and 1 lidar, all with full 360 degree field of view. nuScenes comprises 1000 scenes, each 20s long and fully annotated with 3D bounding boxes for 23 classes and 8 attributes. It has 7x as many annotations and 100x as many images as the pioneering KITTI dataset. We define novel 3D detection and tracking metrics. We also provide careful dataset analysis as well as baselines for lidar and image based detection and tracking. Data, development kit and more information are available online.

nuScenes: A multimodal dataset for autonomous driving

In this paper we describe a low-cost and low-power consumption messaging system based on LoRa technology. More that one billion people worldwide cannot access even the most basic connectivity services. For them even simple messaging services would be of great help, for example to farmers wishing to know the price of goods they want to sell or buy before deciding whether a possibly long, expensive and exhausting trip is undertaken. LoRa networks allow for very long wireless links that can connect villages and towns. This system falls in the category of community networks, where users build their own network where no commercial infrastructure is available. In addition to the simple messaging application, LoRa can be used to distribute sensor information to communities or to provide disaster alerts or meteorological data.

/pdf/a-low-cost-and-low-power-messaging-system-based-on-the-lora-3pwqmm02tc.pdf

A Low-Cost and Low-Power Messaging System Based on the LoRa Wireless Technology

Performance status of the Malawi TVWS pilot as of December 2013 is presented. Basic performance metrics like throughput, latency, SNR have been analyzed using known path loss empirical models like Hata, Asset and Friis. Additionally, this paper presents new data mining tools developed using Python and Perl which were deployed at each TVWS station's ALIX board running on Voyage Linux to abstract useful data for computation of network latency and throughput. Typically, for the longest tested link at 7.5 km, an average SNR = 24.7 dB, data-rate of 420 kbps and latency of 118 ms were observed using the collected data. Empirically, Hata and Asset resolved an average path loss value equal to 155 dB which was 51 dB higher than the ideal Friis Free Space Path Loss (FSPL) at UHF CH31 (554 MHz), BS antenna height of 23 m, receiver antenna height of 2m, and farthest station distance of 7.5 km. These results prove a 2.6 times superior propagation performance of the TVWS network over the commercial fixed broadband wireless network (assuming 2 Mbps backhaul, and all other parameters controlled). These measurements were done in dry season. Results for rainy season will be studied from December 2013 to June 2014.

Malawi Television White Spaces (TVWS) Pilot Network Performance Analysis

This document presents a taxonomy of a set of "Alternative Network
Deployments" that emerged in the last decade with the aim of bringing
Internet connectivity to people or providing a local communication
infrastructure to serve various complementary needs and objectives.
They employ architectures and topologies different from those of
mainstream networks and rely on alternative governance and business
models. The document also surveys the technologies deployed in these
networks, and their differing architectural characteristics, including
a set of definitions and shared properties. The classification
considers models such as Community Networks, Wireless Internet Service
Providers (WISPs), networks owned by individuals but leased out to
network operators who use them as a low-cost medium to reach the
underserved population, networks that provide connectivity by sharing
wireless resources of the users, and rural utility cooperatives.

/pdf/alternative-network-deployments-taxonomy-characterization-4in22r5w27.pdf

Alternative Network Deployments: Taxonomy, Characterization, Technologies, and Architectures

Wireless sensor networks (WSNs) are composed of a large number of sensor nodes that are deployed at target locations. Topology control (TC) is one of the significant fundamental challenges in WSNs because of node energy and computing power constraints. TC algorithms try to produce reduced topology by preserving network connectivity. This study presents a novel TC algorithm based on binary Grey wolf optimisation. It uses the active and inactive schedules of sensor nodes in binary format as well as introduces fitness function to minimise the number of active nodes (ANs) for achieving the target of lifetime expansion of the nodes and network. The proposed algorithm is compared with other TC algorithms. The result reduces a minimum of 10% of ANs and energy consumption by 6.84%. The proposed approach also gives maximum coverage and connectivity. The designed fitness function also benefits in the process of selecting a node with low residual energy to join the active topology. The standard deviation in the remaining energy for the proposed algorithms is lower than the other TC schemes.

https://ietresearch.onlinelibrary.wiley.com/doi/epdf/10.1049/iet-wss.2018.5169

Binary grey wolf optimisation-based topology control for WSNs

An increasing number of implementations of IoT for development use the LoRaWAN protocol as many of them leverage the free network and application servers provided by The Things Networks (TTN) to fulfill their needs. Unfortunately, in some countries in Sub-Saharan Africa and South Asia, Internet access cannot be taken for granted, therefore, TTN might not be available. Moreover, low-cost and low-power consumption options devices are the most sustainable ones. In this paper, we propose a LoRaWAN network with autonomous base stations that can work without Internet connectivity for essential services, while being able to provide additional features whenever Internet access becomes available, even in an intermittent fashion. Security and privacy are preserved, with support for mobile nodes.

https://www.mdpi.com/1999-5903/11/4/93/pdf

Marco Zennaro

Papers

A Low-Cost and Low-Power Messaging System Based on the LoRa Wireless Technology

Malawi Television White Spaces (TVWS) Pilot Network Performance Analysis

Alternative Network Deployments: Taxonomy, Characterization, Technologies, and Architectures

Binary grey wolf optimisation-based topology control for WSNs

A Smart Cities LoRaWAN Network Based on Autonomous Base Stations (BS) for Some Countries with Limited Internet Access