C# Code Example
This chapter provides a detailed explanation of the C# code examples included in the DaoAI World SDK.
Importing Libraries
In the C# example, the following libraries are imported, with DaoAI.DeepLearningCLI being the library used to import the DaoAI World SDK.
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Drawing;
using System.Threading.Tasks;
using DaoAI.DeepLearningCLI;
Reading an Image
The model prediction function in the DaoAI World SDK requires the image to be represented as a one-dimensional array (1D array). Here is the code for reading an image from a file:
// Test image
String root_directory = System.IO.Directory.GetCurrentDirectory();
System.Drawing.Bitmap image = new
System.Drawing.Bitmap("C:\\Users\\daoai\\Downloads\\DW_SDK\\DW_SDK Example\\Data\\maskrcnn_data\\daoai_1.png"); //Image file path
System.Drawing.Bitmap image_copy = new System.Drawing.Bitmap(image);
byte[] pixels = new byte[image.Width * image.Height * 3];
for (int i = 0; i < image.Height; i++)
{
for (int j = 0; j < image.Width; j++)
{
System.Drawing.Color color = image.GetPixel(j, i);
pixels[(i * image.Width + j) * 3] = (byte)(color.R);
pixels[(i * image.Width + j) * 3 + 1] = (byte)(color.G);
pixels[(i * image.Width + j) * 3 + 2] = (byte)(color.B);
}
}
Here, a deep copy of the image is made, and the image object is initialized using the DaoAI.DeepLearningCLI.Image function for further use:
DaoAI.DeepLearningCLI.Image img = new DaoAI.DeepLearningCLI.Image(image.Height, image.Width, DaoAI.DeepLearningCLI.Image.Type.RGB, pixels);
DaoAI.DeepLearningCLI.Image img_copy = img.clone();
byte[] image_data = img_copy.data;
for (int i = 0; i < image.Width; i++)
{
for (int j = 0; j < image.Height; j++)
{
int index_r = i * image.Height + j;
int index_g = i * image.Height + j + image.Width * image.Height;
int index_b = i * image.Height + j + 2 * image.Width * image.Height;
byte r = image_data[index_r];
byte g = image_data[index_g];
byte b = image_data[index_b];
System.Drawing.Color color = Color.FromArgb(r, g, b);
image_copy.SetPixel(i, j, color);
}
}
Loading a Deep Learning Model
String data_path = "..\\..\\..\\..\\Data\\";
String model_path = data_path + "model.dwm";
// Initialize model
DaoAI.DeepLearningCLI.Vision.KeypointDetection model = new DaoAI.DeepLearningCLI.Vision.KeypointDetection(model_path);
Note that each detection task corresponds to a specific object:
// Instance segmentation
DaoAI.DeepLearningCLI.Vision.InstanceSegmentation model(model_path) = new DaoAI.DeepLearningCLI.Vision.InstanceSegmentation(model_path);
// Keypoint detection
DaoAI.DeepLearningCLI.Vision.KeypointDetection model(model_path) = new DaoAI.DeepLearningCLI.Vision.KeypointDetection(model_path);
// Image classification
DaoAI.DeepLearningCLI.Vision.Classification model(model_path) = new DaoAI.DeepLearningCLI.Vision.Classification(model_path);
// Object detection
DaoAI.DeepLearningCLI.Vision.ObjectDetection model(model_path) = new DaoAI.DeepLearningCLI.Vision.ObjectDetection(model_path);
// Unsupervised defect detection
DaoAI.DeepLearningCLI.Vision.UnsupervisedDefectSegmentation model(model_path) = new DaoAI.DeepLearningCLI.Vision.AnomalyDetection(model_path);
// Supervised defect detection
DaoAI.DeepLearningCLI.Vision.SupervisedDefectSegmentation model(model_path) = new DaoAI.DeepLearningCLI.Vision.SemanticSegmentation(model_path);
// OCR
DaoAI.DeepLearningCLI.Vision.OCR model(model_path) = new DaoAI.DeepLearningCLI.Vision.OCR(model_path);
// Positioning model (available only in the industrial version)
DaoAI.DeepLearningCLI.Vision.Positioning model(model_path) = new DaoAI.DeepLearningCLI.Vision.Positioning(model_path);
// Error detection (available only in the industrial version)
DaoAI.DeepLearningCLI.Vision.PresenceChecking model(model_path) = new DaoAI.DeepLearningCLI.Vision.PresenceChecking(model_path);
Using a Deep Learning Model for Prediction
Here, the confidence threshold (CONFIDENT_THRESHOLD) is set to 0.5, and the model.inference() function is called to perform inference with the model. The results are then printed as a JSON using the .toJSONString() method:
Dictionary<DaoAI.DeepLearningCLI.PostProcessType, object> post_params = new Dictionary<DaoAI.DeepLearningCLI.PostProcessType, object>();
post_params[DaoAI.DeepLearningCLI.PostProcessType.CONFIDENT_THRESHOLD] = 0.5;
Console.WriteLine(model.inference(img, post_params).toJSONString());
Post-processing Parameters
The model's predictions can accept post-processing parameters:
DaoAI.DeepLearningCLI.PostProcessType.CONFIDENCE_THRESHOLD:
The confidence threshold filters out results with a confidence lower than the specified value.
post_params[DaoAI.DeepLearningCLI.PostProcessType.IOU_THRESHOLD]:
The IOU threshold filters out results with an IOU lower than the specified value.
post_params[DaoAI.DeepLearningCLI.PostProcessType.SENSITIVITY_THRESHOLD]:
Used in unsupervised defect segmentation (anomaly detection) models to control the sensitivity to defects. The higher the sensitivity, the more defects the model will detect, but it may result in more false positives.
Dictionary<DaoAI.DeepLearningCLI.PostProcessType, object> post_params = new Dictionary<DaoAI.DeepLearningCLI.PostProcessType, object>();
post_params[DaoAI.DeepLearningCLI.PostProcessType.CONFIDENCE_THRESHOLD] = 0.5; // Filters out results with confidence lower than the set value
post_params[DaoAI.DeepLearningCLI.PostProcessType.IOU_THRESHOLD] = 0.5; // Filters out results with IOU lower than the set value
Console.WriteLine(model.inference(img, post_params));
Retrieving Prediction Results
Box (Bounding Box)
The Box represents the bounding box of the predicted results from the model. You can retrieve it as follows:
// Retrieve prediction results
DaoAI.DeepLearningCLI.Vision.KeypointDetectionResult prediction = model.inference(daoaiImage);
// Access bounding box coordinates
double x1 = prediction.boxes[0].x1(); // Top-left corner X-coordinate
double y1 = prediction.boxes[0].y1(); // Top-left corner Y-coordinate
double x2 = prediction.boxes[0].x2(); // Bottom-right corner X-coordinate
double y2 = prediction.boxes[0].y2(); // Bottom-right corner Y-coordinate
Mask (Contour)
Mask provides the contour information of the target object in the prediction. You can extract the vertices of the polygonal region as follows:
// Retrieve prediction results
DaoAI.DeepLearningCLI.Vision.KeypointDetectionResult prediction = model.inference(daoaiImage);
// Extract the first vertex of the polygonal region
double x = prediction.masks[0].toPolygons()[0].points[0].X; // X-coordinate of the vertex
double y = prediction.masks[0].toPolygons()[0].points[0].Y; // Y-coordinate of the vertex
`masks[0]`: Extracts the mask of the first target object.
`toPolygons()`: Converts the mask into polygonal contours.
`points[0]`: Retrieves the first vertex of the polygon.
By connecting all the vertices sequentially, you can form the complete contour of the target object. This method is suitable for scenarios that require detailed boundary information, such as region analysis or fine-grained annotations.
Visualization Output
Generate and save a visualization image that overlays the prediction results on the original image:
DaoAI.DeepLearningCLI.Image result = DaoAI.DeepLearningCLI.Utils.Visualize(img, prediction);
result.save(rootDirectory + "daoai_output.png");
Example of Returned Results
Below is an example of the results returned after running an instance segmentation model prediction.
This result shows the predicted quantity, label names, confidence, bounding boxes, and polygon masks.
{
"Number of detections": 1,
"Detections": [
{
"Label": "zheng",
"Confidence": 0.9523001313209534,
"Box": [
955.1925659179688, 316.0162048339844, 1064.072021484375,
426.4408264160156
],
"Mask": [
[990.0, 316.0],
[988.0, 318.0],
[987.0, 318.0],
[985.0, 320.0],
[982.0, 320.0],
[980.0, 322.0],
[979.0, 322.0],
[974.0, 327.0],
[972.0, 327.0],
[972.0, 328.0],
[1040.0, 316.0]
]
}
],
"ImageHeight": 1200,
"ImageWidth": 1920
}