O‘ZBEKISTON RESPUBLIKASI
RAQAMLI TEXNOLOGIYALAR VAZIRLIGI
MUHAMMAD AL‑XORAZMIY NOMIDAGI
TOSHKENT AXBOROT TEXNOLOGIYALARI UNIVERSITETI
“VIDEOMA’LUMOTLARGA RAQAMLI ISHLOV BERISH”
fanidan
№ 7 - labaratoriya ishi
HISOBOTI
Guruh: 811-20
Talaba: Abdisalomov Avazbek
O‘qituvchi: Xayrullayev Alisher
Toshkent – 2024
7-labaratoriya ishi
Rasmlarni zichlashtirish natijalarini baholash
Dastur kodi:
from skimage.metrics import structural_similarity as ssim
import cv2
import numpy as np
import math
from matplotlib import pyplot as plt
def mse(imageA, imageB):
err = np.sum((imageA.astype("float") - imageB.astype("float")) ** 2)
err /= float(imageA.shape[0] * imageA.shape[1])
return err
def compare(img1, img2):
img1 = cv2.cvtColor(img1, cv2.COLOR_RGB2GRAY)
img2 = cv2.cvtColor(img2, cv2.COLOR_RGB2GRAY)
MSE_val = mse(img1, img2)
SSIM_val = ssim(img1, img2)
RMSI_val =math.sqrt(MSE_val)
if RMSI_val!=0:
PSNR_val=20*math.log10(255/RMSI_val)
RMSI_val=255/pow(10.0,(PSNR_val/20.0))
else:
PSNR_val=100
return MSE_val,SSIM_val,RMSI_val,PSNR_val
org_img = cv2.imread("/kaggle/input/my-imagee/my-image.jpg")
org_img=cv2.cvtColor(org_img, cv2.COLOR_BGR2RGB)
fig, ax = plt.subplots(nrows=10,ncols=2,figsize=(10, 40))
h,w,c=org_img.shape
#inter_type=cv2.INTER_NEAREST
#inter_type=cv2.INTER_LINEAR
#inter_type=cv2.INTER_CUBIC
#inter_type=cv2.INTER_AREA
inter_type=cv2.INTER_LANCZOS4
for i,j in zip(range(1,20,2),range(10)):
ax[j][0].imshow(org_img)
resized_img=cv2.resize(org_img,(int(w/i),int(h/i)),interpolation=inter_type)
res_h,res_w,_=resized_img.shape
ax[j][0].title.set_text("({}:{})/{}=({}:{})".format(h,w,i,res_h,res_w))
resized_img=cv2.resize(resized_img,(w,h),interpolation=inter_type)
MSE,SSIM,RMSI,PSNR=compare(org_img,resized_img,)
ax[j][1].imshow(resized_img)
ax[j][1].title.set_text("MSE:{:.2f} SSIM:{:.2f}RMSE:{:.2f} PSNR:{:.2f}".format(MSE,SSIM,RMSI,PSNR))
plt.show()
Kaggle ilovasida bajarilgan ish havolasi: https://www.kaggle.com/code/avazbekabdisalomov/7-labaratory
Xulosa
Ushbu laboratoriya ishida, rasmlarni zichlashtirish va natijalarini baholash uchun kaggleda dastur yaratdik. Dastur rasmlarni zichlashtirish va ularga mos holda SSIM, MSE, RMSE va PSNR kabi xususiyatlarni hisoblash uchun skimage, cv2, numpy va math kabi kutyaptirlar ishlatilgan. Natijalar, matplotlib kutubxonasidan foydalanib tasvirlar bilan ko'rsatildi.
Asosiy dastur funksiyalari quyidagicha:
mse(imageA, imageB): Ikkala rasmlar o'rtasidagi o'rtacha kvadrat xatolikni hisoblaydi.
compare(img1, img2): Ikki rasmlarni solishtirib, ularning MSE, SSIM, RMSE va PSNR kabi farqliliklarini hisoblaydi.
Asl rasmlar ochiladi va RGB formatiga o'tkaziladi.
Belgilangan qayta o'lchash nisbatlari va interpolatsiya usulini tanlanadi.
O'lchangan rasm olish uchun yangi rasmlar olish, keyin asal rasmlar bilan solishtiriladi.
Har bir qayta o'lchash holati uchun asl va qayta o'lchangan rasmlar bilan bog'liq metrikalar va tasvirlar yaratiladi.
Bu kodlar, rasmlarni o'zgartirish va ularga mos holda qanday o'zgarishlar kelishini baho qilish uchun foydalanilishi mumkin.
Kaggle ilovasida bajarilgan ish havolasi: https://www.kaggle.com/code/avazbekabdisalomov/7-labaratory
|
