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

Nowadays with the evolution of Internet of Things (IoT), building a network of sensors for measuring data from remote locations requires a good plan considering a lot of parameters including power consumption. A Lot of communication technologies such as WIFI, Bluetooth, Zigbee, Lora, Sigfox, and GSM/GPRS are being used based on the application and this application will have some requirements such as communication range, power consumption, and detail about data to be transmitted. In some places, especially the hilly area like Rwanda and where GSM connectivity is already covered, GSM/GPRS may be the best choice for IoT applications. Energy consumption is a big challenge in sensor nodes which are specially supplied by batteries as the lifetime of the node and network depends on the state of charge of the battery. In this paper, we are focusing on static sensor nodes communicating using the GPRS protocol. We acquired current consumption for the sensor node in different locations with their corresponding received signal quality and we tried to experimentally find a mathematical data-driven model for estimating the GSM/GPRS sensor node battery lifetime using the received signal strength indicator (RSSI). This research outcome will help to predict GPRS sensor node life, replacement intervals, and dynamic handover which will in turn provide uninterrupted data service. This model can be deployed in various remote WSN and IoT based applications like forests, volcano, etc. Our research has shown convincing results like when there is a reduction of −30 dBm in RSSI, the current consumption of the radio unit of the node will double.

https://www.mdpi.com/2078-2489/11/11/524/pdf

GPRS Sensor Node Battery Life Span Prediction Based on Received Signal Quality: Experimental Study

EyA-inspired applications (to "Enhance Your Audience") allow for easy automated recordings in the classroom. They also allow for a fast sharing of large amount of educational contents. We discuss on our two prototype apps: AndrEyA and EyApp for mobile devices running Android OS and iOS, with the ultimate goal of facilitating the recoding of academic lectures by individual students themselves. We show the strong differences in video size outputs with respect to video recordings using iPads and our app, without losing sources of information.

/pdf/eyapp-andreya-free-apps-for-the-automated-recording-of-22h0v68rg3.pdf

EyApp & AndrEyA – Free Apps for the Automated Recording of Lessons by Students

In engineering the pedagogical content of most formative programmes includes a significant amount of practical laboratory hands-on activity designed to deliver knowledge acquisition from actual experience alongside traditional face-to-face classroom based lectures and tutorials; this hands-on aspect is not always adequately addressed by current e-learning platforms.
An innovative approach to e-learning in engineering, named computer aided engineering education (CAEE) is about the use of computer aids for the enhanced, interactive delivery of educational materials in different fields of engineering through two separate but related components; one for classroom and another for practical hands-on laboratory work. The component for hands-on laboratory practical work focuses on the use of mixed reality (video-based augmented reality) tools on mobile devices/platforms. 
This paper presents the computer aided engineering education (CAEE) implementation of a laboratory experiment in micro-electronics that highlights some features such as the ability to closely implement an existing laboratory based hands-on experiment with lower associated costs and the ability to conduct the experiment off-line while maintaining existing pedagogical contents and standards.

/pdf/a-low-cost-implementation-of-an-existing-hands-on-laboratory-21bdsx53v8.pdf

A Low Cost Implementation of an Existing Hands-on Laboratory Experiment in Electronic Engineering

We discuss on an innovative system developed in-house to archive and share scientific lectures carried out using modern presentations (PPT, PDF, animations, etc) or, alternatively, the commonly used traditional chalkboards found in classrooms. The system -named " Enhance your Audience " ( EyA ), allows to widening the audience for such activities with little efforts. The technologies employed are low-cost.

EyA system: Automated audio-video-slide recordings

More than fifty low and medium income countries rely on the sea for their economies. Despite its importance, there is a lack of oceanographic scientific data from these countries due to the high cost of equipment and to the running costs of measurement campaigns. In this paper we present a low-cost drifter based on a LoRa communication platform and tested in a coastal area (Gulf of Trieste). The system, built using lowcost off-the-shelf components, has at least the same performances of drifters that cost about three times as much and its operation does not require any recurring costs. The use of a LoRa allows several drifters to operate simultaneously in the same area, providing a rich and homogeneous database for the statistical post processing. The high transmission rate allows an almost instantaneous position determination, facilitating the drifter recovery for successive reuse, which is essential in developing countries. Received on 03 June 2018; accepted on 14 August 2018; published on 11 September 2018

/pdf/on-the-design-of-a-sustainable-ocean-drifter-for-developing-3cvzet2bse.pdf

Marco Zennaro

Papers

GPRS Sensor Node Battery Life Span Prediction Based on Received Signal Quality: Experimental Study

EyApp & AndrEyA – Free Apps for the Automated Recording of Lessons by Students

A Low Cost Implementation of an Existing Hands-on Laboratory Experiment in Electronic Engineering

EyA system: Automated audio-video-slide recordings

On the design of a sustainable ocean drifter for developing countries