IoU, GIoU, DIoU, CIoU, SIoU loss函数python实现

IoU, GIoU, DIoU, CIoU, SIoU 损失函数实现

论文地址：
IoU: UnitBox: An Advanced Object Detection Network
GIoU：Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regression
DIoU&CIoU: Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression
SIoU: SIoU Loss: More Powerful Learning for Bounding Box Regression

IoU

def IoU(box1, box2):
    b1_x1, b1_y1, b1_x2, b1_y2 = box1
    b2_x1, b2_y1, b2_x2, b2_y2 = box2
    
    xx1 = np.maximum(b1_x1, b2_x1)
    yy1 = np.maximum(b1_y1, b2_y1)
    xx2 = np.minimum(b1_x2, b2_x2)
    yy2 = np.minimum(b1_y2, b2_y2)
    
    w = np.maximum(0.0, yy2 - yy1)
    h = np.maximum(0.0, xx2 - xx1)

    inter = w * h
    IoU = inter/((b1_x2-b1_x1)*(b1_y2-b1_y1) + (b2_x2-b2_x1)*(b2_y2-b2_y1) - inter)
    print("IoU: ", IoU)


if __name__ == "__main__":
    box1 = np.array([100, 100, 210, 210])
    box2 = np.array([150, 150, 230, 220])
    IoU(box1, box2)

GIoU

def GIoU(box1, box2):
    b1_x1, b1_y1, b1_x2, b1_y2 = box1
    b2_x1, b2_y1, b2_x2, b2_y2 = box2
    
    # IOU
    xx1 = np.maximum(b1_x1, b2_x1)
    yy1 = np.maximum(b1_y1, b2_y1)
    xx2 = np.minimum(b1_x2, b2_x2)
    yy2 = np.minimum(b1_y2, b2_y2)
    inter_w = np.maximum(0.0, yy2 - yy1)
    inter_h = np.maximum(0.0, xx2 - xx1)
    inter = inter_w * inter_h
    Union = (b1_x2-b1_x1)*(b1_y2-b1_y1) + (b2_x2-b2_x1)*(b2_y2-b2_y1) - inter

    # GIOU
    C_xx1 = np.minimum(b1_x1, b2_x1)
    C_yy1 = np.minimum(b1_y1, b2_y1)
    C_xx2 = np.maximum(b1_x2, b2_x2)
    C_yy2 = np.maximum(b1_y2, b2_y2)
    C_area = (C_xx2 - C_xx1) * (C_yy2 - C_yy1)

    IOU = inter / Union
    GIOU = IOU - abs((C_area-Union)/C_area)
    print("GIOU:", GIOU)

if __name__ == "__main__":
    box1 = np.array([100, 100, 210, 210])
    box2 = np.array([150, 150, 230, 220])
    GIoU(box1, box2)

DIoU

def DIoU(box1, box2):
    b1_x1, b1_y1, b1_x2, b1_y2 = box1
    b2_x1, b2_y1, b2_x2, b2_y2 = box2
    
    # IOU
    xx1 = np.maximum(b1_x1, b2_x1)
    yy1 = np.maximum(b1_y1, b2_y1)
    xx2 = np.minimum(b1_x2, b2_x2)
    yy2 = np.minimum(b1_y2, b2_y2)
    inter_w = np.maximum(0.0, xx2 - xx1)
    inter_h = np.maximum(0.0, yy2 - yy1)
    inter = inter_w * inter_h
    Union = (b1_x2 - b1_x1)*(b1_y2 - b1_y1) + (b2_x2 - b2_x1)*(b2_y2 - b2_y1) - inter

    # DISTANCE
    C_xx1 = np.minimum(b1_x1, b2_x1)
    C_yy1 = np.minimum(b1_y1, b2_y1)
    C_xx2 = np.maximum(b1_x2, b2_x2)
    C_yy2 = np.maximum(b1_y2, b2_y2)
    C_area = (C_xx2 - C_xx1) * (C_yy2 - C_yy1)

    center_b_x = (b1_x1+b1_x2)/2
    center_b_y = (b1_y1+b1_y2)/2
    center_gtb_x = (b2_x1+b2_x2)/2
    center_gtb_y = (b2_y1+b2_y2)/2

    Distance = (center_gtb_x-center_b_x)**2 + (center_gtb_y-center_b_y)**2

    IOU = inter/Union
    DIOU = IOU - Distance/(C_area**2)
    print("DIOU:", DIOU)

if __name__ == "__main__":
    box1 = np.array([100, 100, 210, 210])
    box2 = np.array([150, 150, 230, 220])
    DIoU(box1, box2)

CIOU

def CIoU(box1, box2):
    b1_x1, b1_y1, b1_x2, b1_y2 = box1
    b2_x1, b2_y1, b2_x2, b2_y2 = box2

    # IOU
    xx1 = np.maximum(b1_x1, b2_x1)
    yy1 = np.maximum(b1_y1, b2_y1)
    xx2 = np.minimum(b1_x2, b2_x2)
    yy2 = np.minimum(b1_y2, b2_y2)
    inter_w = np.maximum(0.0, xx2 - xx1)
    inter_h = np.maximum(0.0, yy2 - yy1)
    inter = inter_w*inter_h
    Union = (b1_x2-b1_x1)*(b1_y2-b1_y1) + (b2_x2-b2_x1)*(b2_y2-b2_y1) - inter
    IOU = inter/Union

    C_xx1 = np.minimum(b1_x1, b2_x1)
    C_yy1 = np.minimum(b1_y1, b2_y1)
    C_xx2 = np.maximum(b1_x2, b2_x2)
    C_yy2 = np.maximum(b1_y2, b2_y2)

    # DISTANCE
    center_b_x = (b1_x1 + b1_x2)/2
    center_b_y = (b1_y1 + b1_y2)/2
    center_gtb_x = (b2_x1 + b2_x2)/2
    center_gtb_y = (b2_y1 + b2_y2)/2
    C_area = (C_xx2-C_xx1)*(C_yy2-C_yy1)
    Distance = (center_gtb_x-center_b_x)**2 + (center_gtb_y-center_b_y)**2
    Distance_area = Distance/C_area**2

    # aspect ratio
    pred_w = b1_y2 - b1_y1
    pred_h = b1_x2 - b1_x1
    gt_w = b2_y2 - b2_y1
    gt_h = b2_x2 - b2_x1
    v = (4/(np.pi)**2)*(np.arctan(gt_w/gt_h) - np.arctan(pred_w/pred_h))**2
    alpha = v/((1-IOU) + v)

    CIOU = IOU - Distance_area - alpha*v
    print("CIOU:", CIOU)

if __name__ == "__main__":
    box1 = np.array([100, 100, 210, 210])
    box2 = np.array([150, 150, 230, 220])
    CIoU(box1, box2)

SIoU

def SIoU(box1, box2):
    b1_x1, b1_y1, b1_x2, b1_y2 = box1
    b2_x1, b2_y1, b2_x2, b2_y2 = box2

    # IOU
    xx1 = np.maximum(b1_x1, b2_x1)
    yy1 = np.maximum(b1_y1, b2_y1)
    xx2 = np.minimum(b1_x2, b2_x2)
    yy2 = np.minimum(b1_y2, b2_y2)
    inter_w = np.maximum(0.0, xx2 - xx1)
    inter_h = np.maximum(0.0, yy2 - yy1)
    inter = inter_w*inter_h
    Union = (b1_x2-b1_x1)*(b1_y2-b1_y1) + (b2_x2-b2_x1)*(b2_y2-b2_y1) - inter
    IOU = inter/Union

    center_b_x = (b1_x1 + b1_x2)/2
    center_b_y = (b1_y1 + b1_y2)/2
    center_gtb_x = (b2_x1 + b2_x2)/2
    center_gtb_y = (b2_y1 + b2_y2)/2

    # ANGLE
    sigma = np.sqrt((center_gtb_x-center_b_x)**2 + (center_gtb_y-center_b_y)**2)
    lambda_ch = max(center_gtb_y, center_b_y) - min(center_gtb_y, center_b_y)
    lambda_x = lambda_ch/sigma
    angle = 1 - 2*(np.sin(np.arctan(lambda_x)-np.pi/4)**2)

    # DISTANCE
    lambda_cw = max(center_gtb_x, center_b_x) - min(center_gtb_x, center_b_x)
    Rho_x = ((center_gtb_x-center_b_x)/lambda_cw)**2
    Rho_y = ((center_gtb_y-center_b_y)/lambda_ch)**2
    gamma = 2-angle
    Delat = (1-np.exp(-1*gamma*Rho_x)) + (1-np.exp(-1*gamma*Rho_y))

    # SHAPE
    Theta = 4
    pred_w = b1_y2 - b1_y1
    pred_h = b1_x2 - b1_x1
    gt_w = b2_y2 - b2_y1
    gt_h = b2_x2 - b2_x1 
    Omega_w = abs(pred_w-gt_w)/max(pred_w, gt_w)
    Omega_h = abs(pred_h-gt_h)/max(pred_h, gt_h)
    Omega = (1-np.exp(-1*Omega_w))**Theta + (1-np.exp(-1*Omega_h))**Theta

    SIOU = 1 - IOU + (Delat + Omega)/2
    print("SIOU:", SIOU)

if __name__ == "__main__":
    box1 = np.array([100, 100, 210, 210])
    box2 = np.array([150, 150, 230, 220])
    SIoU(box1, box2)