Kedis ku xidhan hagaha · Dhexdhexaad Heerka

ESRGAN iyo GAN Kediska Xallinta Sare

Baro nashqada kor u qaadaysa sawirka ESRGAN, ganacsigeeda dhabta ah ee garashada, iyo xadka tafaasiilkeeda.

Wadooyinka haga ee la xidhiidhaEsrgan iyo GAN Super Resolution
Su'aal 1 ee 8

Waa maxay 'GAN' ee ku jira ESRGAN?

Horudhac su'aal kasta8 su'aalo

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

  1. Waa maxay 'GAN' ee ku jira ESRGAN?
    • Shabakadda Adversarial Generative
    • Shabakadda Accumulation Network
    • Qiyaasta Gaussian Node
    • Shabakadda Feejignaanta Caalamiga ah
  2. Waa maxay dhismaha uu koronto-dhaliyaha ESRGAN isticmaalo?
    • Unugyada xusuusta muddada-gaaban
    • Hadhaaga-gudbaha cufan (RRDB)
    • Lakabyada xajinta kaabsoosha
    • Transformer blocks is-fiirsashada
  3. Maxay qorayaasha ESRGAN uga saareen wax-soo-saarka dufcadda caadiga ah?
    • Waxa ay aad u hakisay tababarka
    • Waxay kordhisay xusuusta xad dhaafka ah
    • Waxay u janjeertay inay soo bandhigto farshaxan-muuqaal
    • Waxay ahayd mid aan ku habboonayn GAN-yada
  4. Waa maxay ganacsiga-ka-noqoshada ugu weyn ee ESRGAN ka dhigaysa marka la barbar dhigo hababka khasaare-pixel-kaliya?
    • Xawaare ka badan saxnaanta
    • Saxnaanta midabka ee xallinta
    • Xusuus hoose oo ka badan xallinta sare
    • Xaqiiqda garashada ee ka sarreeya daacadnimada pixel saxda ah
  5. Halkee lagu qiyaasaa luminta ESRGAN (VGG)?
    • On sifooyinka ka hor hawlgelinta
    • On pixels RGB cayriin kaliya
    • Marka la eego dhibcaha ugu dambeeya ee takooraha
    • Ku saabsan gelinta Fourier-la beddelay
  6. Waa maxay hal-abuurnimada uu soo bandhigay takooriyaha ESRGAN?
    • Atoencoder oo kali ah oo kontorol ah
    • Midab-takoor is barbar dhig ah oo isbarbar dhigaya xaqiiqada dhabta ah iyo tan dhabta ah ee been abuurka ah
    • Qaabka abaalmarinta xoojinta-barashada
    • Kalasaar hore oo barafaysan
  7. Which competition did the ESRGAN paper report winning?
    • The ImageNet classification challenge
    • The Vesuvius Challenge
    • The PIRM 2018 perceptual super-resolution challenge
    • The COCO image-captioning challenge
  8. Why should generated detail from ESRGAN not be treated as recovered photographic evidence?
    • The model records the original sensor data exactly
    • GANs can only enlarge images without changing pixels
    • The output is guaranteed to match the lost high-resolution image
    • The model can synthesize plausible textures that were not present in the low-resolution input