[Inspection Technique] DeepSky Feature 2 – Auto Annotation
Introducing the practical features of the AI image inspection software 'DeepSky' from a field perspective!
▼Feature: Auto Annotation This is a function that automatically executes annotations. It was previously a beta feature, but it has become an official feature starting from Ver. 2.2.0.0. Many of you may know that annotations are very important in object recognition. However, when there are many target objects in an image, the amount of work required for annotation increases, and as the workload increases, mistakes such as forgetting to annotate or making errors inevitably rise. If annotations are forgotten, DeepSky adjusts parameters to determine that "this is something that should not be found (even though it looks similar) for the same object," which can significantly lower the recognition rate of the object, leading to various issues. (See Figure 1: Annotation Forgetting) This is where the auto annotation feature comes into play. By using auto annotation, with just a click of a button, it automatically suggests annotations like "Based on the previous training data, you would want to annotate here, right?" (See Figures 2, 3, and 4)
basic information
After that, if there are any areas that need to be confirmed and corrected by the operator, they can add or remove annotations or adjust the size of the boxes, and then it will be complete. (See Figures 5 and 6) This allows the operator to be freed from the simple task of arranging a large number of boxes and to focus on the important task of verifying the automatically generated annotations. While performing auto-annotation requires some prior training, it can significantly reduce the workload of annotating a large number of training images before training. In such cases, manual annotation can be done on the first few images to begin training. After that, auto-annotation can be applied to the remaining images to automate the rest of the annotations. If you are thinking, "Annotation is a hassle" → "I should have someone else do it ^^;", please try using the "Auto Annotation" feature. You will find it to be a very simple and convenient function.
Price range
Delivery Time
Applications/Examples of results
▼Function: Data Augmentation (Inflation) As a side note, please also use the "Data Augmentation" (Inflation) feature. When used in conjunction with auto-annotation, it can automatically generate a large number of images and annotation data from a small number of training images. We encourage you to try out this practical feature from a field perspective.
