{"id":50302,"date":"2019-12-06T00:00:00","date_gmt":"2019-12-06T00:00:00","guid":{"rendered":"https:\/\/www.techopedia.com\/computer-vision-revolutionizing-research-in-2020-and-beyond\/"},"modified":"2019-11-26T15:55:08","modified_gmt":"2019-11-26T15:55:08","slug":"computer-vision-revolutionizing-research-in-2020-and-beyond","status":"publish","type":"post","link":"https:\/\/www.techopedia.com\/computer-vision-revolutionizing-research-in-2020-and-beyond\/2\/34115","title":{"rendered":"Computer Vision: Revolutionizing Research in 2020 and Beyond"},"content":{"rendered":"
Computer vision<\/a> is expected to prosper in the coming years as it's set to become a $48.6 billion industry by 2022<\/a>. Organizations are making use of its benefits in improving security, marketing, and production efforts. Engineers (and scientists, too), firmly believe there are more advantageous applications to be expected from the technology in the coming years.<\/p>\n So does computer vision actually make research activities more efficient and effective? And, how exactly will the technology be applied and what potential does it really have?<\/p>\n The following provides several examples of how computer vision is being utilized in various industries and how it can help further advance studies in respective fields.<\/p>\n According to studies<\/a>, 90% of medical data is image-based. This makes computer vision a perfect fit in understanding more about an illness based on its texture, color, and shape, combined with previous knowledge about the disease.<\/p>\n Bone imaging is needed for identifying the possible occurrence of cancer, infections, arthritis, and dysplasia. A computed tomography (CT) scan is usually done to create a 3D representation of the bones and areas of interest.<\/p>\n However, since a large portion of the bones is made of porous, spongy material, the developed model from X-ray attenuation can be inconclusive. This can be corrected through manual editing, but this is an exhaustingly long process. This not only makes the procedure expensive but also harder for doctors and researchers to gather time-sensitive information.<\/p>\n With the help of deep learning<\/a> and computer vision, it’s possible to produce bone segments from CT scans with speed and accuracy. (Read A Tour of Deep Learning Models<\/a>.)<\/p>\n The resulting image also has better intensity and contrast, making it easier for doctors and researchers to distinguish healthy cells.<\/p>\n While the technology is still being developed and perfected, it will soon be easier to detect brain tumors, cancer cells, and bone defects without having to go through a long and costly process.<\/p>\n Sports biomechanics: the physics of sports. By utilizing the benefits of computer vision, data for studying sports biomechanics becomes more detailed and is easier to obtain.<\/p>\n Second Spectrum, which is the official partner of the NBA responsible for analyzing sports data, uses a combination<\/a> of machine learning<\/a>, big data<\/a>, and computer vision. Using these processes, they were able to deconstruct the aspects of rebounding including the most frequent rebound location, the average rebound height, the rebound success rate, and the shot locations relative to the completion of an offensive or defensive rebound.<\/p>\n There’s also a study about detecting swimming strokes from beginning to end<\/a>. With the help of computer vision and machine learning, researchers were able to conduct discrete event detection and locate the exact frame when the swim stroke starts and ends. This study can help athletes improve the timing and span of their strokes in order to maximize speed and minimize drag.<\/p>\n Gait analysis<\/a> can also be conducted in conjunction with motion capture to study locomotion and muscle activity. With the help of cameras and computer vision, it becomes easier to calculate trajectory, deduce movement patterns, and measure muscle activity in three dimensions.<\/p>\n Human beings count many things, from stars in galaxies, to icebergs in the ocean, to cells under a microscope. Counting absorbs many man-hours of labor but can be done much more efficiently by a computer vision system trained by machine learning. (Read Job Role: Machine Learning Engineer<\/a>.)<\/p>\n Cell counting is an age-old scientific practice. The hemocytometer was invented in the 19th century France was developed to make counting easier, and although a piece of glass etched with grids could allow a doctor to determine that a patient has a low red cell count, contemporary science has an exploding need for faster methods.<\/p>\nProcess Images Faster and Better<\/span><\/h2>\n
Detecting Motion with High Accuracy<\/span><\/h2>\n
Counting with Machines Instead of Humans<\/span><\/h2>\n