Kuiz Inpainting dan Outpainting

Jawapan dan penjelasan muncul semasa anda mengambil kuiz.

1. Apa yang dilakukan dalam mengecat?
   • Menjana semula kawasan bertopeng di dalam imej
   • Memanjangkan imej melepasi sempadannya
   • Menukar teks kepada pertuturan
   • Memampatkan fail imej
2. Apakah yang dilakukan oleh lukisan luar?
   • Mengeluarkan warna daripada foto
   • Menghasilkan kandungan baharu di luar tepi asal
   • Mengesan muka
   • Menajamkan piksel kabur
3. How does RePaint use the known region during diffusion-based inpainting?
   • It permanently freezes every unmasked pixel with no resampling
   • It removes the mask before denoising
   • It repeatedly resamples and conditions generation on known pixels during reverse diffusion
   • It trains a new model from scratch for each image
4. What architecture does LaMa introduce for large-mask image inpainting?
   • A language-only transformer
   • A nearest-neighbor search over captions
   • A fixed-color fill with no learned model
   • Fast Fourier Convolutions with a receptive field that supports image-wide context
5. Apakah yang membimbing kandungan yang muncul di kawasan bertopeng?
   • Gesaan teks dan piksel sekeliling
   • Hanya nama fail
   • Jam peranti
   • Tahap mampatan imej
6. Mengapa cat luar berguna untuk pembingkaian semula?
   • Ia memadamkan subjek
   • Ia boleh menukar nisbah bidang foto dengan menambahkan pemandangan
   • Ia menyulitkan imej
   • Ia mengurangkan resolusi
7. Besides Fourier Convolutions, what two components does the LaMa paper identify in its large-mask inpainting method?
   • A speech decoder and image captioner
   • A fixed image palette and nearest-neighbor lookup
   • A high receptive-field perceptual loss and large training masks
   • A video tracker and optical-flow classifier
8. What does a mask specify in an inpainting task?
   • The image’s camera model
   • The order of training examples
   • The output file’s copyright status
   • Which image region is to be modified or synthesized