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

Ubiquitous Wireless Sensor Networks (UWSN) has emerged as a most powerful and efficient tool for monitoring, assessing and managing the environmental pollution related issues particularly in getting the real time data from widely spread locations, to a centralized location, which is not possible otherwise. Wireless Sensor Networks has also been widely used by military, security agencies, industries, healthcare monitoring systems and even for managing, to some extent, the manmade/ natural disasters. During the past few years, the state of Himachal Pradesh, located in the Western Himalayan region of India, has made remarkable progress in developing new industrial areas with growing population along traffic startups having a good network of transport infrastructure. Such a growth in new industrial establishments has opened up growing opportunities for the people living in this area, but at the same time a significant increase in the pollution level has also been noticed in the environmental air as well as in water posing a serious threat to the habitants. This conceptual paper presents a solution for monitoring air pollution to acquire real time values in the state of Himachal Pradesh, India. A successful implementation of Ubiquitous WSN and its connectivity with the Himachal State Wide Area Network (HIMSWAN) will help in analyzing air pollution parameters in a better way by collecting the real time data from the widely spread locations at a central location, analyzing it and then generating the forewarning for the people of the state.

Ubiquitous Wireless Sensor Networks for Environmental Monitoring in the Western Himalayan Region of India

Connectivity is key for IoT and smart cities. Unfortunately, a stable Internet connection is scarce in developing countries. LoRaWAN standalone base station solutions can be used to fill the gaps. But since these difficulties may not affect everyone, then, affordable wireless communication, such as Wi-Fi, with direct access to Internet from the collection node, may be useful for data transmission. This article, first, discusses a coverage study based on LoRaWAN autonomous base stations and, then, extends the architectural model proposed in [3] to take into account the Wi-Fi protocol, thus diversifying the implementation choices. A gateway (Wi-IoT) capable of providing Wi-Fi access, on the one hand, and collecting, processing and monitoring data as a mini-server, on the other hand, will be proposed as proof of concept. From the node to the gateway, data will be compressed and sent securely. A user who connects to Wi-IoT will, then, be able to access his data.

A LoRaWAN Coverage Testbed and a Multi-optional Communication Architecture for Smart City Feasibility in Developing Countries

Environmental monitoring is of growing interest around the globe and there is a clear need for better education in this area. However, the progress is incredibly fast here and wireless sensor networking is a relatively new area of research, so no standard curricula are available yet. In 2012, the International School “Sustainable Wireless ICT Solutions for Environmental Monitoring” was organized by the International Center for Theoretical Physics (ICTP) in collaboration with a few other entities. The school aimed at exposing young scientists from around the world to the newest wireless solutions for environmental monitoring. The school was followed by an awareness conference, which aimed at identifying issues that would benefit from a wider scientific collaboration. This paper begins with general information about the ICTP educational activities related to wireless technologies and applications. Then it focuses on the program of the school and conference. It shows what topics were included (and what were disregard) and what approaches were applied in order to achieve the best results with the (very) limited resources available. Details are omitted here as the complete materials of the both events are available at the Internet (http://wireless.ictp.it) under the Creative Commons. Keywords—DIGI, Fundacion Escuela Latinoamericana de Redes, EsLaRed, International Center for Theoretical Physics, National Institute of Telecommunications, Network Startup Resource Center, NIT, wireless sensor networks, WSN.

/pdf/educating-on-wireless-solutions-for-environmental-monitoring-40g3w8ycnn.pdf

Educating on Wireless Solutions for Environmental Monitoring

Although LoRa (long range) technology is designed to tackle interference and power consumption issues, interference from other LoRa and non-LoRa networks limits its performance. In practice, LoRa performance also gets constrained by multipath fading. This paper presents the study of LoRa transmission over Rayleigh fading. We considered different scenarios of LoRa networks along with simultaneous parallel transmissions, change in transmission bandwidths, and Doppler shifts. Results demonstrate deviation in the BER because of multipath and interference. It is also evident that even for a change in transmission bandwidth or Doppler shift for Rayleigh fading channel, LoRa performance gets impacted.

LoRa Transmission Over Rayleigh Fading Channels in Presence of Interference

With the advancements in the Internet of Things (IoT), machine-to-machine communication, big-data, and the associated environment, a new model of the Industrial Internet of Things (IIoT) has emerged. The IIoT brings sensors, intelligent machines and tools, instruments, and analytics together for applications like manufacturing, robotics, and many others. Most of these applications require adaptive and autonomous behavior, Quality of Service (QoS), efficient resource allocation, and reservation. One of these crucial challenges is to build and maintain a data communication schedule. Time Slotted Channel Hopping (TSCH) based network operation promises required QoS for low-power applications and enables high reliability. 6TiSCH layer is being developed by standardizing the protocol stack to achieve industrial performance requirements by using IPv6 over IEEE802.15.4e TSCH MAC. 6TiSCH aims to manage the schedule and configure it with the topology and traffic requirements in the industrial environment. This paper proposes a novel low latency autonomous scheduling scheme for the 6TiSCH networks. It generates a segmented schedule for the network where all source nodes can send application data packets to the root node in a single slotframe. The performance of the proposed technique is compared with existing scheduling techniques. Our scheme outperforms the other techniques. The result shows that the latency is reduced up to 41% in comparison with the other scheduling schemes. The proposed scheme has a lower radio duty cycle as the node’s ON time is reduced, making it more energy efficient and reliable.

Marco Zennaro

Papers

Ubiquitous Wireless Sensor Networks for Environmental Monitoring in the Western Himalayan Region of India

A LoRaWAN Coverage Testbed and a Multi-optional Communication Architecture for Smart City Feasibility in Developing Countries

Educating on Wireless Solutions for Environmental Monitoring

LoRa Transmission Over Rayleigh Fading Channels in Presence of Interference

6TiSCH Low Latency Autonomous Scheduling for Industrial Internet of Things