引導式測驗 · 簡單 等級
修復和修復測驗
了解蒙版如何指導影像修復以及修復如何將影像內容擴展到原始邊框和畫布限制之外。
問題 1 的 8
補漆有什麼作用?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- 補漆有什麼作用?
- 重新產生影像內的遮罩區域
- 將圖像擴展到其邊界之外
- 將文字轉換為語音
- 壓縮圖像文件
- 外畫有什麼作用?
- 去除照片中的顏色
- 產生超出原始邊緣的新內容
- 偵測人臉
- 銳利化模糊像素
- RePaint 在基於擴散的修補過程中如何使用已知區域?
- 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
- LaMa 為大遮蔽影像修補引入了什麼架構?
- 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
- 是什麼引導出現在屏蔽區域的內容?
- 文字提示和周圍的像素
- 僅檔案名稱
- 設備的時鐘
- 影像的壓縮級別
- 為什麼外畫對於重構很有用?
- 它刪除了主題
- 它可以透過添加風景來改變照片的縱橫比
- 它對圖像進行加密
- 它會降低分辨率
- 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
- 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