Inspection
- Kristyna Janku
Owned by Kristyna Janku
Inspection shows how the input image is processed for each module, where detected objects are displayed for each image, providing detailed information such as: rectangle coordinates, ID, image size, confidence percentage, etc.
The Context (application variable), is displayed in JSON format for each module. The data will be available in Inspection after you activate a model for any module within the Flow.
Module-independent data
Image Dimension (Pixels)
{
"globalData": null,
"image": {
"type": "<numpy>",
"shape": [
837,
1305,
3
]
}
Anomaly Detector - JSON Data
Detected Rectangle Data - XY Coordinates, Dimension[px], Area[px], ID, Color, Class Name
"detectedRectangles": [
{
"x": 1086,
"y": 510,
"width": 3,
"height": 3,
"area": 2,
"id": 1603111484990003,
"classNames": [
{
"color": "#ff00ff",
"color_bgr": [
255,
0,
255
],
"id": 1603111688057,
"label": "Scratch"
}
....Heatmap Data - Dimension, Color, ID & Class Name
"heatmaps": [
[
{
"type": "<numpy>",
"shape": [
768,
1024,
1
]
},
{
"color": "#ff0000",
"color_bgr": [
0,
0,
255
],
"id": 1,
"label": "Defect"
}
]
....Classifier - JSON Data
Detected Rectangle Data - Class Names, ID & Confidence % (Accuracy)
Accuracy * 100 = Accuracy %
Detected Rectangle for Product A has 99.99% confidence
Surface Detection - JSON Data
Detected Rectangle Data - XY Coordinates, Dimension, Area, ID, Color, Class Name
Heatmap Data - Dimension, Color, ID & Class Name
Detector - JSON Data
Detected Rectangle Data - Coordinates, Dimension, ID, Class Name, Confidence Percentage
Accuracy * 100 = Accuracy %
OCR - JSON Data
Detected Rectangle Info - ID, Dimension, OCR Text, Confidence Percentage