Collaborating Authors

Sensing and Signal Processing

Basic Example of Neural Style Transfer – Predictive Hacks


This post is a practical example of Neural Style Transfer based on the paper A Neural Algorithm of Artistic Style (Gatys et al.). For this example we will use the pretained Arbitrary Image Stylization module which is available in TensorFlow Hub. We will work with Python and tensorflow 2.x. Neural style transfer is an optimization technique used to take two images--a content image and a style reference image (such as an artwork by a famous painter)--and blend them together so the output image looks like the content image, but "painted" in the style of the style reference image. This is implemented by optimizing the output image to match the content statistics of the content image and the style statistics of the style reference image.

An Experiment in Deep Learning with Wild Animal Trail Camera Data


Trail cameras are automatically triggered by animal movements. They are used by ecologists and wildlife managers around the world to study wild animal behavior and help preserve endangered species. I want to see if MATLAB image processing and deep learning can be used to identify individual animal species that visit trail cameras. We are going to start with off-the-shelf functionality--nothing specialized for this particular task. My partners on this project are Heather Gorr and Jim Sanderson. Heather is a machine learning expert at MathWorks.

How we built an easy-to-use image segmentation tool with transfer learning


Training an image segmentation model on new images can be daunting, especially when you need to label your own data. To make this task easier and faster, we built a user-friendly tool that lets you build this entire process in a single Jupyter notebook. The main benefits of this tool are that it is easy-to-use, all in one platform, and well-integrated with existing data science workflows. Through interactive widgets and command prompts, we built a user-friendly way to label images and train the model. On top of that, everything can run in a single Jupyter notebook, making it quick and easy to spin up a model, without much overhead.

Models Trained to Keep the Trains Running


Steady advances in machine vision techniques such as convolutional neural networks powered by graphics processors and emerging technologies like neuromorphic silicon retina "event cameras" are creating a range of new predictive monitoring and maintenance use cases. We've reported on several, including using machine vision systems to help utilities monitor transmission lines and towers linked to wildfires in California. Now, AI software vendor Ignitarium and partner AVerMedia, an image capture and video transmission specialist, have expanded deployment an aircraft-based platform for detecting railway track obstructions. The AI-based visual "defect detection" platform incorporates Ignitarium's AI software implemented on Nvidia's edge AI platform used to automatically control onboard cameras. The system is designed to keep cameras focused on the track center during airborne inspections.

Author Correction: Analysis of the Human Protein Atlas Image Classification competition


Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit

Reolink Argus 3 security camera review: New look, same great performance


Reolink found a good niche with its first Argus, a truly wireless home security camera with just the essential features. Now on its third iteration, the camera has a new look and some smart additions, including an integrated spotlight and color night vision. Even better, Reolink has introduced these new twists without mucking with the camera's easy operation and simple feature set. With its modular size, white enclosure, and black face, the redesigned Reolink 3 resembles a more squat Arlo Pro 3. On the front are a status LED, a 230-lumen spotlight (something else it shares with the Arlo Pro 3), six infrared LEDs for night vision, a fixed focal length lens with a 120-degree field of view, daylight sensor, microphone, and a passive-infrared (PIR) sensor for motion detection. On the back are a battery status LED, micro-USB port, and speaker.

EETimes - LeapMind Enters IP Business with AI Accelerator


LeapMind (Tokyo, Japan) announced its entry into the processor IP business with Efficiera, an ultra-low-power AI inference accelerator IP product. Efficiera is optimized for models that have been heavily quantized using LeapMind's'extremely low-bit quantization' software techniques. It is designed for convolutional neural networks (CNNs), the type of network typically used for image processing and analysis tasks today. "This is the company's first hardware IP product. But we are working on the development of a core technology called extreme quantization technology that operates at both software and hardware-IP levels with a network optimized for practical applications and a dedicated compiler," a LeapMind spokesperson told EE Times.

Object Detection with No Data Thanks to Template Matching


How to implement custom object detection with template matching. Today, state-of-the-art object detection algorithms (algorithms aiming to detect objects in pictures) are using neural networks such as Yolov4. Template matching is a technique in digital image processing for finding small parts of an image that matches a template image. It is a much simpler solution than a neural network to conduct object detection. In my experience, combining a neural network like Yolov4 and object detection with template matching here is a good way to considerably improve your neural network performance! When you use OpenCV template matching, your template slides pixel by pixel on your image.

Introduction To Image Denoising


Image enhancement is an important research topic in image processing and computer vision. It is mainly used as image pre-processing or post-processing to make the processed image clearer for subsequent image analysis and understanding. There are many sources of noise in images, and these noises come from various aspects such as image acquisition, transmission, and compression. The types of noise are also different, such as salt and pepper noise, Gaussian noise, etc. There are different processing algorithms for different noises.

Bonds between AAAS and Hubble span three decades


On New Year's Day, 1925, Henry Russell, director of the Princeton University Observatory, presented to the joint meeting of the American Association for the Advancement of Science and the American Astronomical Society a research paper that would change humanity's understanding of the Universe and our place in it. Busy with ongoing observations at the Mount Wilson Observatory near Los Angeles, the paper's author, Edwin Hubble, mailed his work to the meeting rather than traveling to Washington, D.C., to present it himself. Using Mount Wilson's 100-inch telescope—the world's largest from 1917 to 1949—Hubble had taken images of the Andromeda Nebula, determining that it is a galaxy of its own, composed of individual stars. The observation proved that our Milky Way does not take up nearly the entire Universe, as many astronomers of the time believed, but instead is one galaxy of many. “This paper is the product of a young man of conspicuous and recognized ability in a field which he has made peculiarly his own,” Russell and Joel Stebbins, secretary of the American Astronomical Society, wrote in a letter to the AAAS Committee on Awards. “It has already expanded one hundred-fold the known volume of the material universe.” Of approximately 1,700 scientists who shared research at the meeting, the committee chose two—Hubble and a zoologist named L. R. Cleveland—to share the 1924 Thousand Dollar Prize, an award honoring the most noteworthy contributions to science presented at the AAAS Annual Meeting each year. Now called the Newcomb Cleveland Prize, the yearly award goes to the author or authors of a particularly impactful paper published in Science . “The Association Prize will never be awarded to a more appreciative student than your present choice,” Hubble wrote to AAAS secretary Burton Livingston in February 1925. “The occasion of the reward, however, must be regarded as a triumph of modern instruments rather than a personal achievement.… This was accomplished by the use of the largest telescopes in existence.” Decades later, the National Aeronautics and Space Administration launched into Earth orbit an instrument exponentially larger and more precise than anything available to Hubble, and it bore his name. In April 1990, the Hubble Space Telescope became the first major optical telescope to be placed in space. Free from the atmospheric distortion and light pollution faced by land-based telescopes, Hubble has provided data leading to more than 17,000 peer-reviewed publications on topics including star formation, galaxy mergers, and dark matter, the largely unseen mass that occupies most of the Universe. NASA is celebrating a series of the Hubble mission's 30th-anniversary milestones this year, from the launch in April to its “first light” image in May and its observations of Supernova 1987A in August. Throughout the mission's first three decades, AAAS has honored its scientists with awards and fellowships and used their expertise to guide AAAS programs. AAAS and Hubble also overlap in their vision for the science, technology, engineering, and mathematics (STEM) enterprise, as both work to enhance public engagement, educate and diversify the STEM workforce, and foster international collaboration. “Hubble, slam dunk, aligns with almost all of them,” astronomer Kathryn Flanagan said of AAAS's goals. Prior to retiring in March, Flanagan oversaw the Hubble mission as deputy director and interim director of the Baltimore-based Space Telescope Science Institute (STScI), a nonprofit science center operated for NASA and responsible for Hubble's science operations and public outreach. She also chaired AAAS's astronomy section in 2016 and was inducted as an elected AAAS fellow this year, the latest of dozens of Hubble-affiliated scientists to receive the honor. “Hubble was, when I assessed it a few years back, contributing as a recommended or required component in half the state departments of education in the U.S.,” Flanagan said, referencing the prevalence of the telescope's findings in public school curricula. “It reached half the public middle school students in the country.” Flanagan also highlighted the Hubble mission's dedication to international collaboration through its partnership with the European Space Agency in development and operations, publicly available archives, research time allotted to scientists from around the world, and a double-anonymous peer-review process, which ensures that unconscious bias does not play a role in determining which research proposals are selected. Many AAAS programs, including Science in the Classroom, the Center for Science Diplomacy, and SEA Change, have a similar focus on science education, international collaboration, and achieving equity in STEM disciplines. Carol Christian, Hubble Space Telescope outreach project scientist at STScI, works to share the mission's discoveries with the world. Christian is part of the Hubble news team, highlighting science results from hundreds of refereed papers each year. Additionally, she leads a program that brings Hubble data to individuals with blindness and visual impairments using 3D printing. Christian also has served as a screener for the AAAS Kavli Science Journalism Awards since 2010. Each year, approximately 70 scientists volunteer to assess the scientific accuracy of the award program's print, digital, and video reporting candidates. Christian said that the review panels consist of researchers from diverse disciplines, including geology and the life sciences. “You learn a lot from the people that are your fellow reviewers,” she said. Prior to working with the Kavli Awards, Christian brought her expertise in satellite imagery to the U.S. Department of State, where she worked from 2003 to 2007 through the AAAS Science and Technology Policy Fellowships program, which places more than 250 scientists each year in policy roles across all three branches of the federal government. She enjoyed putting her scientific knowledge to practical use during her time as a fellow. “Suddenly, to be doing something that actually has import—policy may depend on it, economics may depend on it—that was pretty interesting,” Christian said. Margaret Burbidge, an influential astrophysicist who died in April at the age of 100, was perhaps the first scientist to contribute to both the Hubble mission and AAAS. During her term as president of AAAS in 1983, she was on the team that developed the Faint Object Spectrograph, one of the five science instruments on Hubble when it launched. Burbidge and her colleagues designed the instrument to detect the physical and chemical properties of faint objects in galaxies beyond our own, and it provided the first strong observational evidence for the existence of a supermassive black hole in the core of another galaxy. “She was one of the names that we all knew of as pioneers in the field of modern astronomy,” astrophysicist Jennifer Wiseman said following Burbidge's death. “An inspirational person as a brilliant scientist and in particular as a courageous woman scientist who opened up the field for many of us coming later on.” Like Burbidge did for a time, Wiseman works with the Hubble Space Telescope and is also active in AAAS activities. As Hubble's senior project scientist, Wiseman works with a team of scientists and engineers at NASA's Goddard Space Flight Center to keep the telescope as scientifically productive as possible. Prior to her NASA career, Wiseman served as an American Physical Society congressional science fellow, a role under the umbrella of the AAAS Science and Technology Policy Fellowships. She brought her subsequent connection with Hubble to a broad AAAS audience by co-organizing a symposium on the Hubble Space Telescope mission—its scientific discoveries, image processing feats, and future promise—at the AAAS 2015 Annual Meeting, and she has given Hubble overview talks at subsequent AAAS gatherings. Since 2010, Wiseman has also directed the AAAS Dialogue on Science, Ethics, and Religion (DoSER) program, which works to foster communication between scientific and religious communities. “Being an active scientist while attentive to public science engagement helps me to recognize how scientists and the other communities that DoSER works with—including ethicists and religious communities—can better understand each other and our common interests in science, technology, and their positive contributions to society,” Wiseman said. “Astronomy in general gives everyone a sense of humility, awe, wonder, and curiosity,” she added. “I enjoy sharing with many types of public audiences the amazing things we're discovering with telescopes and other scientific facilities. I'm very humbled and encouraged by how such discoveries can inspire deeper discussions of meaning and purpose and of using science to help one another, as we all ask questions together of who we are as human beings in an awesome Universe and how we can use our knowledge to uplift the human spirit.”