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

Artificial intelligence (AI) will likely be an instrumental part of progress towards the United Nations’ Sustainable Development Goals (SDGs). However, its adoption and impact are limited by the immense power consumption, strong connectivity requirements and high costs of cloud-based deployments. TinyML is a new technology that allows machine learning (ML) models to run on low-cost, low-power microcontrollers, circumventing many of these issues. We believe that TinyML has a significant role to play in achieving the SDGs and facilitating scientific research in areas such as environmental monitoring, physics of complex systems and energy management. To broaden access and participation and increase the impact of this new technology, we present an initiative that is creating and supporting a global network of academic institutions working on TinyML in developing countries. We suggest the development of additional open educational resources, South – South academic collaboration and pilot projects of at-scale TinyML solutions aimed at addressing the SDGs.

TinyML: Applied AI for Development

Relatively low-cost noise sensors can be utilized to detect for aberrant powerline noise, which is an early indicator and warning of potential power reliability and stability issues. By taking preemptive action, power outages may be avoided. Accordingly, the practicality as pertains to the utilization of low-cost noise sensors, segueing to scalability and extensibility, is examined. The strategic placement of sensors at key distribution poles has been examined previously. This paper examines how low-cost noise sensors provide an opportunity for comprehensive coverage and more likely detection of certain powerline noise aberrations at the "edge" against a set of compiled heuristics.

Detecting Powerline Noise with Low-Cost Noise Sensors for Power Outage Mitigation

Medical products are very sensitive to temperature; the improper temperature may lead to their inefficacity. Apart from products that are stored at room temperature, remaining medical products are stored in electronically controlled refrigerators. A lot of researchers have proposed different refrigeration systems controlled with the help of the internet of things (IoT). Due to some issues such as storage capacity, computing energy, and computing speed, data processing in IoT-based applications is generally done at the cloud through cloud computing technology. Those applications are suffering issues like latency, data control, internet connectivity, network traffic, and operation cost. In this paper, we are experimentally developing a four rooms fridge controlled with an Arduino board that embeds a machine learning (ML) algorithm to control the temperature for efficient storage of medical products. We tried to develop an ML model that will monitor the closing and opening of the fridge door (while taking some medicines), predict and display the remaining time for the internal temperature to go beyond the acceptable temperature range. The result from our experiments shows that the model runs onto the controller and can predict well the internal fridge temperature at an accuracy of 96%.

Development of a TinyML based four-chamber refrigerator (TBFCR) for efficiently storing pharmaceutical products: Case Study: Pharmacies in Rwanda

An Analysis of many IoT deployments showed that most of them can address the sustainable development goals (SDGs) and the UN’s 2030 agenda. Interestingly, most of these projects concentrate on five SDGs: 1) industry, innovation, infrastructure; 2) smart cities and communities; 3) affordable and clean energy; 4) good health and well-being; and 5) responsible production and consumption. Examples include a remote water-monitoring solution that ensures clean water in regions with an indigenous population and smart lighting initiatives in Chinese cities that halve total power output. 

Guest Editorial Special Issue on Sustainable Solutions for the Internet of Things

Connectivity is the essential requirement for a smart city, and traditional techniques are too costly to connect a great number of devices which might be served with lower data throughput, but require longer ranges to accommodate the requirements of IoT. Although cellular technologies and Wi-Fi have proved an enormous success in connecting billions of users to the Internet, they are not the best to connect objects with limited throughput requirements, limited processing power and low power consumption requirements. These needs are being addressed by Low Power Wide Area Networks (LPWAN) that use unlicensed frequencies and by newer cellular protocols aimed at lowering the cost of the devices by limiting their bandwidth capabilities in the licensed bands. This paper discusses these technologies and the choice made to provide an IoT infrastructure for the city of Maputo in Mozambique.

Marco Zennaro

Papers

TinyML: Applied AI for Development

Detecting Powerline Noise with Low-Cost Noise Sensors for Power Outage Mitigation

Development of a TinyML based four-chamber refrigerator (TBFCR) for efficiently storing pharmaceutical products: Case Study: Pharmacies in Rwanda

Guest Editorial Special Issue on Sustainable Solutions for the Internet of Things

Evaluation of IoT gateways for developing communities: smart Maputo