Computer Vision for Industrial Engineers - Continuous Improvement

Adam Bennett

fromMatroid

28 Dec, 2022

Matroid Computer Vision | SEE IT. DETECT IT.

Deep-Learning technology helps industrial engineers with continuous improvement, line balancing, and optimizing manual labor.

artificial intelligence

- computer vision

- digital manufacturing

- machine learning

- manufacturing systems

Industrial engineers are key to ensuring production lines don’t miss a beat. Their eyes are always on the lines looking for every potential weak link within the production flow. Ideally an industrial engineer could continuously keep an eye on all things at once to detect imbalances on production lines that cause decreased yields and poor utilization of manual labor, all of which diminish productivity and profitability.

By applying video cameras in areas that depend on manual labor, industrial engineers can feed that digital data into a Computer Vision AI system and begin to capture video analytics on cycle time, non-value add time, anomalies, action sequences, and much more. The Computer Vision AI models are trained to understand various actions and sequences of those actions so that manual labor operations are digitally validated and recorded. Intelligent Computer Vision AI systems integrate into other various IT/OT systems, such as machines, MES, ERP, etc., bringing actionable insight to operational systems and digital traceability to business systems.

A continuous capture of manual labor provides valuable information about cycle times and non-value add time spent during operational hours. The power of deep-learning is applied to differentiate the amount of time spent between effective production time vs. non-value add time. With this information further conclusions can be drawn, and alerts can be created. For example, if too much time is spent assembling a part this can be flagged for further analysis. The industrial engineer can view a video playback of the operation to determine if this is a result of something that can be resolved (i.e. operational inefficiency, quality defect, or true anomaly). If an operator takes more time than what has been normalized to assemble parts, Computer Vision AI can help capture the root cause of common quality defects. In other scenarios information is captured over time that can provide clues to any non-productive trends. This type of information allows industrial engineers to apply their lean-manufacturing skillsets for continuous improvement, such as reengineering workspaces, tooling, and improving line balancing upstream or downstream.

Humans have 5 senses and of those, 80% of all impressions come from sight. Visual data is one of the easiest sources of data to generate. Computer Vision AI leverages the latest in machine-learning / deep-learning technologies to make sense of visual data. Using any type of camera technology (visible and invisible spectrums) Computer Vision AI learns and detects what is visible from the data and provides analytics on the detections. These new tools are empowering industrial engineering experts to continuously improve their operations to keep a competitive edge with quality, efficiency, and productivity.

Search for articles and topics on Wevolver

"artificial intelligence"

"computer vision"

"digital manufacturing"

"machine learning"

"manufacturing systems"

"3d printing"

Explore topics

Interactive StylEx system aims to unveil how a classifier operates, and help root out bias

Designed to fight bias in both neural network classifiers and the datasets on which they are trained, StylEx lets you play with sliders to figure out exactly what makes a model tick.

Gareth Halfacree

02 Feb, 2022

artificial intelligence

- computer vision

- machine learning

Neural network technology can help solve some of the toughest problems facing humanity today, but it comes with one major caveat: It’s a black box, taking in data and spitting out results without ever being able to explain how or why. As neural networks are deployed in increasing numbers, their impact grows accordingly — as does the potential for bias to wreak havoc on those who feel its impact.

This is where a team from Google Search, Hebrew University, MIT, Tel Aviv University, and the Weizmann Institute of Science come in with a potential answer for classification networks: StylEx, which provides “disentangled attributes” designed to explain exactly why a given image was classified in a certain way — figure out, effectively, what makes a cat a cat or a car a car, at least from the perspective of the classifier itself.

Rooting out bias via explanations

The problem of bias in machine learning is well understood. Researchers at North Carolina State University and Pennsylvania State University highlighted issues with algorithmic bias in artificially intelligent hiring platforms nearly two years ago; in 2019 a team at Carnegie Mellon University proposed a system for looking inside machine learning models to find influential proxies; and team comprised of computer scientists from Princeton and Stanford Universities proposed tackling the problem in the data sets themselves.

The approach taken for StylEx is different, and considerably more hands-on. StylEx, its creators explain, provides a means of training a model for a generative adversarial network (GAN) to create a “StyleSpace” with explicable attributes specific to the classifier. “These can then be used to visualize the effect of changing multiple attributes per image,” the team writes in the paper’s abstract, “thus providing image-specific explanations.”

Those explanations aren’t presented as a simple list of properties: Key to StylEx is

Image: By placing images on a virtual "rubber sheet," this ML approach needs little training.

One-short learning approach, inspired by human intuition, works with just one training image

Designed to dramatically reduce the amount of training data needed for an image recognition system, this one-shot approach "inspired by nativism" takes a leaf from humans' ability to intuit and abstract.

Gareth Halfacree

04 Feb, 2022

artificial intelligence

- computer vision

- machine learning

The human ability to generalize information, using abstraction to recognize everything from the shapes of letters and numbers to abstract doodles based on very limited experience, may be within reach of machine learning systems thanks to work by researchers at the Universities of Chicago and Illinois at Urbana-Champaign.

In a paper dubbed “Learning from One and One One Shot,” the research team details an approach “inspired by nativism” which allows a character recognition system to achieve human-like accuracy based on as little as one example image per class — by imagining the images, suitably processed and simplified, as being printed on a rubber sheet which can be stretched, squeezed, and bent into a range of shapes.

Abstracted learning

Machine learning systems are powerful, but suffer from one major drawback: The need for extensive training, expensive both computationally and in the datasets they ingest. A leading object recognition model may, for example, be trained on tens or even hundreds of thousands of images which have to be sourced and processed.

Approaches which look to dramatically reduce the size of training datasets are commonly referred to as “few-shot learning;” those, like this work from Haizi Yu, Igor Mineyev, Lav R. Varshney, and James Evans, which go still further to being able to generalize a single example of a given class are thus “one-shot.”

It’s a powerful idea, and one which draws inspiration from the human brain — specifically, its capacity for abstraction. “Our mind freely conceives more equivalent examples,” the researchers explain, “ignoring low-level variations in between. Babies can differentiate things quickly based on general appearance, i.e., how things ‘look in general’ (topologically, geometrically, structurally, etc.) recognizing that they may be translated, rotated, scaled, or deformed.”

Applying images to a "rubber sheet" provides a means of quantifying thei

Computer Vision for Industrial Engineers - Continuous Improvement

Deep-Learning technology helps industrial engineers with continuous improvement, line balancing, and optimizing manual labor.

Search for articles and topics on Wevolver

Rooting out bias via explanations

Abstracted learning

HyperTransformer bests rivals at few-shot learning - and takes advantage of unlabeled samples

Autonomous drones, triggered by a machine learning algorithm, could solve the urban feral pigeon problem

Artificial intelligence deployed to enhance sports performance