728x90 ๋ฐ์ํ generative model3 cGAN/Pix2Pix 1. GAN 2. cGAN ์ด๋ค ์ซ์๋ฅผ ๋ง๋ค์ด๋ผ์ง์ ๋ํ ์ ๋ณด๋ฅผ ๋ฃ์ด์ฃผ๋ ๊ฒ(์ด๋ค ํด๋์ค์ ํด๋นํ๋์ง์ ๋ํ ์ ๋ณด) ex) 7์ ๋ง๋ค๊ณ ์ ํ๋ค๋ฉด condition vector์ 7์ ๋ฃ์ด์ฃผ๊ณ , z(noise)์๋ ๋๋คํ๊ฒ ์ํ๋งํด์ 7์ด๋ผ๋ ์๋ฏธ๋ฅผ ๊ฐ์ง๋ ๋๋คํ ํํ๋ฅผ ๋ง๋ค์ด์ค 3. Pix2Pix image to image translation : ์ด๋ฏธ์ง์ ํน์ ์์์ ๋ค๋ฅธ ์์์ผ๋ก ๋ฐ๊ฟ์ฃผ๋ ๊ฒ์ ์๋ฏธ ex) ์๊ทธ๋ฆผ → ์ค์ ์ฌ์ง์ผ๋ก translation ์ด๋ฏธ์ง ์์ฒด๋ฅผ condition์ผ๋ก ๋ฐ์๋ฒ๋ฆผ (์ด๋ฏธ์ง ์์ฒด๊ฐ ์ ๋ณด๊ฐ ๋๋ ๊ฒ์) ์ฆ, ์ด๋ฏธ์ง๊ฐ ๋ค์ด์์ ๋, ๊ฑฐ๊ธฐ์ ๋ถํฉํ๋ output์ ํํ๋ก ๋ง๋ค์ด์ค noise vector z๋ฅผ ์์ pixel ์ ์ ๋ ฅ์ผ๋ก ๋ฐ์ pixel์ ์์ธกํจ (์ด๋ฏธ์ง๋ฅผ ์ด๋ฏธ.. 2023. 7. 7. GAN: Generative Adversarial Nets 0. Abstract ๋ณธ๋ฌธ We propose a new framework for estimating generative models via an adversarial process, in which we simultaneously train two models: a generative model G that captures the data distribution, and a discriminative model D that estimates the probability that a sample came from the training data rather than G. The training procedure for G is to maximize the probability of D making a m.. 2023. 7. 6. CycleGAN 0. Abstract Figure 1: Given any two unordered image collections X and Y , our algorithm learns to automatically “translate” an image from one into the other and vice versa: (left) Monet paintings and landscape photos from Flickr; (center) zebras and horses from ImageNet; (right) summer and winter Yosemite photos from Flickr. Example application (bottom): using a collection of paintings of famous.. 2023. 7. 5. ์ด์ 1 ๋ค์ 728x90 ๋ฐ์ํ