Deep learning is becoming a commonly heard word in business, science and programming circles. However, the concept is still quite new to many people. This important type of algorithm is a subset of machine learning. It uses neural networks and pattern recognition to develop deep insights that would ordinarily be indecipherable by other types of programs. Due to increased computing power and strides in deep learning programming, this technology is driving a true technical revolution.
The easiest example of deep learning to understand is visual search algorithms. Imagine a photo of a tree. Now think of the millions of potential photos of trees and thousands of different types of trees. How is the computer supposed to know that they should all be labeled “tree” and be indexed that way in visual search? The answer is deep learning algorithms which input a “training set” of data. The program becomes generally familiar with what a tree looks like. When a new photo of a tree is inserted, it can accurately describe and categorize the image.
While advanced deep learning programs are still in the works, there are many exciting possibilities. For example, many researchers believe deep learning technology could produce ground breaking medical drugs that would cure disease and make humans much healthier. Today, scientists move to slowly with trials and experimentation to try the full range of options. However, a deep learning program may be able to create a realistic simulation that produces amazing results.
Other opportunities for breakthroughs based on deep learning include weather forecasting, quality control in manufacturing (getting mistakes below one in a million), economic models and improving machine vision systems.
No matter what, the steady advance of this technology is already having profound impacts on society and technology.
Artificial intelligence (or AI) has seen a recent surge in use across a wide variety of industries. One type of AI currently being used in image processing is called Deep Learning. Deep learning algorithms use neural networking to process enormous data sets in order to carry out a desired task. The applications of deep learning […]
Deep learning is currently being used in the automotive industry for a number of inspection applications. Most manufacturers have large databases of past material that can easily be used by deep learning algorithms for initial learning. Once the algorithm has been learned, it can be maintained and adjusted over time, just like any other piece […]
Machine Vision and Deep Learning Deep learning is becoming a commonly heard word in business, science and programming circles. However, the concept is still quite new to many people. This important type of algorithm is a subset of machine learning. It uses neural networks and pattern recognition to develop deep insights that would ordinarily be […]
Machine Vision and Deep Learning Artificial intelligence, or AI, has seen a recent surge in use across a wide variety of industries. Anyone who has used Google or Amazon has been subject to one of the more popular uses of AI in the customized advertising. Another popular use of AI is image processing, which has […]
CHALLENGES IN PAD PRINTING INSPECTION Multi-step tampography can lead to considerable but tolerable relative shifts (registration problem) Varying quantities of ink applied results in fonts or lines that are thicker/thinner in appearance Random texture of substrate, like brushed or otherwise decorated metal With Cognex ViDi Suite, the automated aesthetic inspection of complex pad printing […]
CHALLENGES IN MACHINED PART INSPECTION Typically there are many different types of complex shapes Varying surface properties depending on machining tool quality and varying properties of the blank material need to be tolerated Some defects only show under a very particular combination of illumination, camera and surface orientation With ViDi Suite, the automated visual inspection […]
Update: ViDi is now owned and operated by Cognex Corporation Integro Technologies, a premier machine vision integrator, announces its partnership with ViDi Systems SA, an innovative software firm in Switzerland. ViDi is revolutionizing the industry by commercializing the first Deep Learning-based image analysis solution dedicated to the machine vision market. ViDi recently designed and developed […]
CHALLENGES IN AUTOMATIC DETECTION Typical defects like hits, scratches or stains are often hardly discernible at the early production stages because of the parts’ rough and strongly textured surfaces. Normal variations and non-significant anomalies in the material need to be tolerated. Stains & hits often manifest themselves by a local change in contrast caused by […]