Application for evaluation machine
InnerEye Ltd.
inner eye
InnerEye's Sense I is software that uses expert judgment obtained from brain waves to generate image classification AI models. In jobs that require the judgment of experts, such as visual inspections and security in the manufacturing industry, it is difficult and time-consuming to pass on judgment and judgment standards from one person to another, which are difficult to explain in words. Therefore, this product performs AI learning by measuring the brain waves generated when an expert makes a decision and quantifying the criteria for determining what is considered abnormal, and generates an AI that can detect abnormalities. Masu.
Problem solving using brainwave AI
soft label and hard label
Normal AI learning uses hard labels, which determine whether the training data is ``0'' or ``1.''
For example, when creating an image classification AI that can inspect the shipping of tomatoes, label "0" to image data that is not suitable for shipping, and "1" to image data that is OK for shipping.
This type of labeling that clearly indicates whether a product is "defective" or "good" is called a hard label, and is widely used in the original learning method.
However, in actual operation, there are cases where it is not clear whether a value is 0 or 1, or where the criteria for judgment differs depending on the person.
To solve this problem, the learning method used in this product is soft labels.
In contrast to hard labels, labels that contain information between "0" and "1" that constantly fluctuate like brain waves are called soft labels.
If a person reacts strongly to an event that they judge, their brain waves will swing significantly (*event-related potential). This makes it possible to link the brain wave waveform with the image data being viewed at the time and assign a label to it.
In addition, due to the characteristic that brain waves can vary in size, it is possible to decipher how confident a person is in image data, so ambiguous judgments between "0" and "1" can be labeled. It can be treated as.
In shipping inspections, etc., it is difficult to label training data due to ambiguous human judgments, so it is advantageous to use soft labels such as electroencephalogram data.
Click here for SenseI use cases in manufacturing industry
event-related potential
The graph on the right shows the brain wave response when visual information is processed in the visual cortex of the brain after a person views an image.
When classifying a good product as a "non-target image" and a defective product as a "target image", we find that when looking at a non-target image, an electroencephalogram response with a constant amplitude is produced, as shown by the blue line. , when looking at the target image, a very distinctive amplitude appears as a red line. This is called an event-related potential.
This product determines that the image that was being viewed when the brain wave response like the red line appeared is the "target image" and labels it.
EEG measurement demo video
We will introduce a demonstration video at an airport baggage inspection to show how AI learning is performed from the human brain using InnerEye's SenseI.
Please also see measurement result examples.
Measurement result
In the upper right graph of the image below, the tester marks baggage without dangerous goods as "OK" and baggage with dangerous goods as "NG".
The waveform of the brain waves when viewing each image data is displayed.
The green curve is the brain wave waveform when viewing the "OK" image, and the red curve is the brain wave waveform when viewing the "NG" image.
The graph at the bottom right shows the concentration level of the examiner.
I maintain a high level of concentration when I'm sorting baggage, but I can see that my level of concentration drops sharply during the breaks in between.
Images that were displayed when the concentration level was low can be displayed again without being used for classification results.
On the left side, images of baggage that were actually judged as NG are displayed in order of confidence.
Images when the amplitude of brain waves is large and concentration is high are labeled as data with high confidence.
Anomaly detection AI is completed based on labeled data.
In actual operation, automatic classification is possible by loading the image data you want to classify into this AI.
If you are having trouble using AI for visual inspection in the manufacturing industry or at airports, you may be able to solve the problem with brain wave AI!
If you are interested or would like to try it out, please use the link below to request a demonstration or to request information.
Use Case
Various use cases are envisioned, such as manufacturing, airport security, medical sites, and agriculture.
Some of them are introduced below.
[Airport security] Succession of know-how on dangerous goods inspection
The inspection of dangerous goods in baggage inspection requires strict and instantaneous judgment, so it is necessary to accumulate experience.
By measuring the brain waves of experts, it can be used to improve accuracy.
[Medical field] Succession of complex skills required for diagnosis and treatment
Diagnosis of disease requires experience and judgment because it is derived from multiple test results. In addition, it is sometimes difficult to create a manual, such as the exquisite amount of force and delicacy in treatment.
By using brain waves, you can inherit skills that are difficult to verbalize.
[Manufacturing industry] Inheriting know-how in defective product inspection
By measuring the brain waves of skilled workers, it can be used to pass on advanced techniques that are difficult to put into words, such as the "intuition of a craftsman," such as visually inspecting defective products during the manufacturing process.
Click here for use case
[Agriculture] Crop cultivation know-how inheritance
Cultivation requires multiple judgments that include not only the weather, temperature, and humidity, but also crop information such as variety, number of cultivation days, and nutritional value. So-called "craftsmanship", such as a lot of experience and intuition based on it, can be inherited by using the brain waves of experts.
Click here for examples
マクニカは古賀琢麻選手を応援しています
NASCARレーサー古賀琢麻選手とのスポンサー契約を締結しました！
この度のスポンサードを通じ、自社の最先端テクノロジーの実証のため、脳波の実証実験を実施しました。 極限状態のドライバーの脳波を含む生体情報を収集し、集中度、ストレス等を分析し、事故防止、快適な走行を可能にする次世代モビリティ開発に活用できるようにします。
詳細はこちら
Sense Plus EEG analysis platform
Sense Plus EEG analysis platform
It is a platform that analyzes human inner emotions and mental states from the state of the brain and visualizes them as numerical data in time series.
Click here for details
free demonstration
Guidance of the demonstration
Would you like to actually measure brain waves?
What is the "skill transfer of experts" based on brain waves? What is AI that learns human consciousness and judgment?
I think there are a lot of unknown things, so why not try measuring your brain waves first and experience what it's like?
After the actual measurement, we will provide support from proposals to implementation and operation so that we can use AI using EEG at the customer's site while including examples of AI utilization in our activities so far. increase.
Click here for details
Click here for related articles
List of articles
If you are interested in other Braintech-related articles, please click here.
Document download
If you want to know the details, please download the materials from here.
リンク先：https://www.macnica.co.jp/business/ai/manufacturers/innereye/products/135823/
*Customer information registration is required for materials.
Inquiry
For inquiries about InnerEye products, please contact us below.
InnerEye manufacturer information Top
If you want to return to the InnerEye manufacturer information top page, please click below.