How to deal with false positive results in OpenCV?

Question

Python

Guys, I'm trying to set up OpenCV so that I can find the right product in the photo. But I ran into the problem of false positives (supposedly shows the borders of the product in the place where it does not exist).
For example, there is such a photo of a mouse:

Based on it, I make my own haar cascade.
Then, using this code, I try to find the mouse in the photo:

import cv2
import sys

imagePath = sys.argv[1]
cascMouse = 'logitech_cascade.xml'
mouseCascade = cv2.CascadeClassifier(cascMouse)

# Read the image
image = cv2.imread(imagePath)
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

mouses = mouseCascade.detectMultiScale(
    gray,
    scaleFactor=1.1,
    minNeighbors=5,
    minSize=(100, 100),
)

for (x, y, w, h) in mouses:
    cv2.rectangle(image, (x, y), (x+w, y+h), (0, 255, 255), 2)

cv2.imshow("Electronic items detector", image)

I get the following result:

We see false positives. I'm trying to play with the parameters of detectMultiScale()

mouses = mouseCascade.detectMultiScale(
    gray,
    scaleFactor=1.5,
    minNeighbors=10,
    minSize=(100, 100),
)

The result gets better:

Trying to find the same object in another photo:

With the same settings, it falsely detects the area where there is no object and does not detect the area where the same object is.
What is the best way to configure OpenCV? I am sure that I am doing something wrong and most likely I am missing some points.