BioGTK.Transforms Class Reference

Public Member Functions
	Transforms (int target_length)
float[]	ApplyImage (Bitmap b)
PointPromotion	ApplyCoords (PointPromotion org_point, int orgw, int orgh)
BoxPromotion	ApplyBox (BoxPromotion org_box, int orgw, int orgh)

Detailed Description

Definition at line 14 of file Transforms.cs.

Constructor & Destructor Documentation

Transforms()

BioGTK.Transforms.Transforms

(

int

target_length

)

Definition at line 16 of file Transforms.cs.

        {
            this.mTargetLength = target_length;
        }

Member Function Documentation

ApplyBox()

BoxPromotion BioGTK.Transforms.ApplyBox	(	BoxPromotion	org_box,
		int	orgw,
		int	orgh )

The function ApplyBox takes a BoxPromotion object, along with its original width and height, and applies coordinate transformations to return a new BoxPromotion object.

Parameters

BoxPromotion	The BoxPromotion class represents a box with coordinates for the top-left and bottom-right corners.
orgw	The orgw parameter represents the original width of the box.
orgh	The parameter "orgh" represents the original height of the box.

Returns

The method is returning a BoxPromotion object.

Definition at line 148 of file Transforms.cs.

        {
            BoxPromotion box = new BoxPromotion();
 
            PointPromotion left = this.ApplyCoords((org_box as BoxPromotion).mLeftUp, orgw, orgh);
            PointPromotion lefrightt = this.ApplyCoords((org_box as BoxPromotion).mRightBottom, orgw, orgh);
 
            box.mLeftUp = left;
            box.mRightBottom = lefrightt;
            return box;
        }

ApplyCoords()

PointPromotion BioGTK.Transforms.ApplyCoords	(	PointPromotion	org_point,
		int	orgw,
		int	orgh )

The function takes in a point and its original width and height, applies a scaling factor based on a target length, and returns a new point with adjusted coordinates.

Parameters

PointPromotion	PointPromotion is a class that represents a point with additional properties related to promotions. It has properties like X and Y coordinates, and an Optype property.
orgw	The original width of the shape.
orgh	The parameter "orgh" represents the original height of the shape.

Returns

The method is returning a new instance of the PointPromotion class, which is assigned to the variable newpointp.

Definition at line 126 of file Transforms.cs.

        {
            int neww = 0;
            int newh = 0;
            this.GetPreprocessShape(orgw, orgh, this.mTargetLength, ref neww, ref newh);
            PointPromotion newpointp = new PointPromotion(org_point.foreGround);
            float scalx = (float)neww / (float)orgw;
            float scaly = (float)newh / (float)orgh;
            newpointp.X = (int)(org_point.X * scalx);
            newpointp.Y = (int)(org_point.Y * scaly);
 
            return newpointp;
        }

ApplyImage()

float[] BioGTK.Transforms.ApplyImage

(

Bitmap

b

)

The function takes a BioImage object, resizes it, calculates the mean and standard deviation of each channel, and returns a transformed image.

Parameters

BioImage

The BioImage parameter represents an image object that contains the image data. It likely has properties such as SizeX and SizeY, which represent the width and height of the image, respectively.

Returns

The method is returning a float array called "transformedImg".

Definition at line 29 of file Transforms.cs.

        {
            int neww = 0;
            int newh = 0;
            if(SAM.SAM2 && MicroSAM.theSingleton == null)
            {
                neww = 1024;
                newh = 1024;
            }
            else
            this.GetPreprocessShape(b.SizeX, b.SizeY, this.mTargetLength, ref neww, ref newh);
 
            float[,,] resizeImg = this.Resize(b,neww,newh);
 
            float[] means = new float[resizeImg.GetLength(0)];
            for (int i = 0; i < resizeImg.GetLength(0); i++)
            {
                float[] data = new float[resizeImg.GetLength(1) * resizeImg.GetLength(2)];
                for (int j = 0; j < resizeImg.GetLength(1); j++)
                {
                    for (int k = 0; k < resizeImg.GetLength(2); k++)
                    {
                        data[j * resizeImg.GetLength(2) + k] = resizeImg[i, j, k];
                    }
                }
                means[i] = (float)MathNet.Numerics.Statistics.Statistics.Mean(data);
            }
 
            float[] stdDev = new float[resizeImg.GetLength(0)];
            for (int i = 0; i < resizeImg.GetLength(0); i++)
            {
                float[] data = new float[resizeImg.GetLength(1) * resizeImg.GetLength(2)];
                for (int j = 0; j < resizeImg.GetLength(1); j++)
                {
                    for (int k = 0; k < resizeImg.GetLength(2); k++)
                    {
                        data[j * resizeImg.GetLength(2) + k] = resizeImg[i, j, k];
                    }
                }
                stdDev[i] = (float)MathNet.Numerics.Statistics.Statistics.StandardDeviation(data);
            }
 
 
            float[] transformedImg = new float[3 * this.mTargetLength * this.mTargetLength];
            for (int i = 0; i < neww; i++)
            {
                for (int j = 0; j < newh; j++)
                {
                    int index = j * this.mTargetLength + i;
                    transformedImg[index] = (resizeImg[0, i, j] - means[0]) / stdDev[0];
                    transformedImg[this.mTargetLength * this.mTargetLength + index] = (resizeImg[1, i, j] - means[1]) / stdDev[1];
                    transformedImg[2 * this.mTargetLength * this.mTargetLength + index] = (resizeImg[2, i, j] - means[2]) / stdDev[2];
                }
            }
 
            return transformedImg;
        }

---

The documentation for this class was generated from the following file:

• BioGTK/SAM/Transforms.cs