DENKblatt
04/2024NEW: PIXEL INFORMATION FOR ANNOTATION
Our developers have realised another customer request.
The new feature is particularly interesting for use cases in which the object sizes are relevant. (e.g. for sorting or filtering processes)
A new feature is that when annotating in the DENK VISION AI Hub, the user is now shown the area content of the marked objects (in pixels).
For example, objects from/up to a certain size can be reliably marked so that only these are included in the AI training.
The new feature is particularly interesting for use cases in which the object sizes are relevant. (e.g. for sorting or filtering processes)
A new feature is that when annotating in the DENK VISION AI Hub, the user is now shown the area content of the marked objects (in pixels).
For example, objects from/up to a certain size can be reliably marked so that only these are included in the AI training.
TIP OF THE WEEK: SHORT VIDEOS FOCUS AREA
For the "experienced" users of the DENK VISION AI Hub, our developers have provided new functions: Focus & Revision Area.
The function of a "focus area" is helpful in AI training, for example, if there are a large number of objects on an image and it would be too time-consuming to mark each one individually. The customer can define a focus area so that the objects outside the focus area can be "left out". → Saves time when editing the image. Nevertheless, a larger focus area always means better training for the AI model.
Our tip: Use our demo videos (~ 1min.) with illustrations of both functions.
Where? Directly in the annotation area of the DENK VISION AI Hub.
The function of a "focus area" is helpful in AI training, for example, if there are a large number of objects on an image and it would be too time-consuming to mark each one individually. The customer can define a focus area so that the objects outside the focus area can be "left out". → Saves time when editing the image. Nevertheless, a larger focus area always means better training for the AI model.
Our tip: Use our demo videos (~ 1min.) with illustrations of both functions.
Where? Directly in the annotation area of the DENK VISION AI Hub.
USE CASE: WASHERS
How long would you need? What makes the washers different?
Although our eyes can be very fast, speed and tirelessness are two unbeatable strengths of an image recognition AI. The AI detects differences in quality or surface finish, as with these metal components, in milliseconds. (see hardware examples)
👍 CPU i5 8th Gen: 170 ms on average
👍 👍 internal GPU intel UHD 620: 160 ms on average
👍 👍 👍 GPU NVIDIA GeForce RTX 3080: < 90 ms.
What is this used for? For all production chains for sorting, positioning or quality control, among other things.
In which areas? Can be used in all areas - in industry, agriculture, the pharmaceutical and healthcare sectors, waste management and many more.
