profile
I'm a Graduate student studying Deep Learning.KR๐Ÿ‘จโ€๐Ÿ’ป
post-thumbnail

Progressive Pose Attention Transfer for Person Image Generation

์ด ๋…ผ๋ฌธ์€ ํฌ์ฆˆ ์ „๋‹ฌ์„ ์œ„ํ•œ ์ƒˆ๋กœ์šด GAN ๋„คํŠธ์›Œํฌ๋ฅผ ์ œ์•ˆํ•ฉ๋‹ˆ๋‹ค. ์ฆ‰, ์ฃผ์–ด์ง„ ์‚ฌ๋žŒ์˜ ํฌ์ฆˆ๋ฅผ ํƒ€๊ฒŸ ํฌ์ฆˆ๋กœ ์ „๋‹ฌํ•˜๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค. Generator๋Š” pose-attentional transfer ๋ธ”๋Ÿญ์œผ๋กœ ๊ตฌ์„ฑ๋˜์–ด ํŠน์ • ๊ด€์‹ฌ ์˜์—ญ์„ ์ „์†กํ•˜์—ฌ ์‚ฌ๋žŒ ์ด๋ฏธ์ง€๋ฅผ ์ ์ง„์ ์œผ๋กœ ์ƒ์„ฑํ•ฉ๋‹ˆ

2021๋…„ 8์›” 3์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€
post-thumbnail

PoNA : Pose-guided Non-local Attention for Human Pose Transfer

๋ณธ ๋…ผ๋ฌธ์€ ๋‹จ์ˆœํ™”๋œ ๊ณ„๋‹จ์‹(cascaded) ๋ธ”๋ก์ด ์ ์šฉ๋œ ์ƒ์„ฑ์  ์ ๋Œ€ ๋„คํŠธ์›Œํฌ(GAN)๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ์ƒˆ๋กœ์šด ์‚ฌ๋žŒ์˜ ํฌ์ฆˆ ์ „์†ก ๋ฐฉ๋ฒ•์„ ์ œ์•ˆํ•˜๊ณ , ๊ฐ ๋ธ”๋ก์—์„œ ์ „์†กํ•  ์ด๋ฏธ์ง€ ํŠน์ง•์˜ ๋” ์ค‘์š”ํ•œ ์˜์—ญ์„ ์„ ํƒํ•˜๊ธฐ ์œ„ํ•ด long-range dependency ๋ฐฉ์‹์„ ์‚ฌ์šฉํ•˜๋Š” Po

2021๋…„ 8์›” 2์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€

XingGAN for Person Image Generation

์ด ๋…ผ๋ฌธ์€ ํฌ์ฆˆ์™€ ์ด๋ฏธ์ง€์˜ ์ƒ๊ด€๊ด€๊ณ„๋ฅผ ํฌ์ฐฉํ•˜์—ฌ ์ƒˆ๋กœ์šด ํฌ์ฆˆ์˜ ๋”ฐ๋ฅธ ์ด๋ฏธ์ง€ ์ƒ์„ฑ์„ ์ •ํ™•ํ•˜๊ฒŒ ํ•˜๋Š” ์ƒˆ๋กœ์šด ๋ฐฉ๋ฒ•์„ ์„ค๋ช…ํ•ฉ๋‹ˆ๋‹ค.

2021๋…„ 7์›” 30์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€
post-thumbnail

K ์ตœ๊ทผ์ ‘ ์ด์›ƒ ์•Œ๊ณ ๋ฆฌ์ฆ˜ (KNN)

K-Nearest Neighbor(KNN) ์ƒˆ๋กœ์šด ๋ฐ์ดํ„ฐ๊ฐ€ ์ฃผ์–ด์กŒ์„ ๋•Œ, ๊ฐ€์žฅ ๊ฐ€๊นŒ์šด K๊ฐœ์˜ ์ด์›ƒ์˜ ์ •๋ณด๋กœ ์ƒˆ๋กœ์šด ๋ฐ์ดํ„ฐ๋ฅผ ์˜ˆ์ธกํ•˜๋Š” ๋ฐฉ๋ฒ• ๋ถ„๋ฅ˜ : ๋‹ค์ˆ˜๊ฒฐ๋กœ ์˜ˆ์ธก KNeighborsClassifier ํšŒ๊ท€ : ํ‰๊ท  ๊ฐ’์œผ๋กœ ๊ฒฐ๊ณผ ์˜ˆ์ธก KNeighborsRegressor

2021๋…„ 6์›” 25์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€

ํ•  ์ผ

์ž์œจ์ฃผํ–‰ ํ”„๋กœ์ ํŠธ ์‹œ๋‚˜๋ฆฌ์˜ค ๋ถ„๋ฅ˜ ๋…ผ๋ฌธ ๋ณด๊ธฐ 4๋‹จ๊ณ„๊นŒ์ง€ ์„ธ๋ถ€๋‚ด์šฉ ์ƒ๊ฐํ•ด๋ณด๊ธฐ ๊ณต๋ถ€ ๊ด€๋ จ ์—ฐ๊ตฌ๋…ผ๋ฌธ ๋ณด๊ธฐ(ํฌ์ฆˆ ๋ง๊ณ  ๋‹ค๋ฅธ๊ฒƒ๋„ ํฌํ•จ) ๋ฒจ๋กœ๊ทธ ์ฑ„์šฐ๊ธฐ ํŒŒ์ด์ฌ์œผ๋กœ ์—ฌ๋Ÿฌ ๊ฐ€์ง€ ํ”„๋กœ๊ทธ๋žจ ์งœ๋ณด๊ธฐ ์ฝ”๋“œ ํ”„๋ ˆ์†Œ ๊ณต๋ถ€ ํŒŒ์ดํ† ์น˜ ๋ฐฐ์šฐ๊ธฐ, ์ผ๋‹จ ํ…์„œํ”Œ๋กœ์šฐ๋Š” ๋ณด๋ฅ˜(?) ๋ฐ”๋กœ ์•„์นด๋ฐ๋ฏธ ๋…นํ™” ํ† ์ต

2021๋…„ 6์›” 23์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€
post-thumbnail

Binary Search VS Linear Search

์–ด๋–ค ์ž๋ฃŒ๊ตฌ์กฐ๋ฅผ ์–ธ์ œ ์„ ํƒํ•˜๋Š”์ง€ ์ค‘์š”ํ•œ ๊ฒƒ ์ฒ˜๋Ÿผ ๋ฌด์Šจ ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ์–ธ์ œ ์„ ํƒํ•˜๋Š”์ง€ ๋ฐฐ์›Œ์•ผ ํ•ฉ๋‹ˆ๋‹ค.

2021๋…„ 6์›” 21์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€
post-thumbnail

NumPy-2

๋‹ค์–‘ํ•œ ์ฐจ์›์˜ ๋ฐฐ์—ด์„ ํšจ์œจ์ ์œผ๋กœ ๋‹ค๋ฃฐ์ˆ˜ ์žˆ๋Š” ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ

2021๋…„ 6์›” 20์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€
post-thumbnail

Disentangled Person Image Generation

์ด ๋…ผ๋ฌธ์€ ํฌ์ฆˆ, ์ „๊ฒฝ, ๋ฐฐ๊ฒฝ์„ ๊ฐ๊ฐ ๋ถ„๋ฆฌํ•ด์„œ embedding feature๋ฅผ ํ•™์Šต์‹œํ‚จ ๋‹ค์Œ, ๊ฐ™์ด ๋””์ฝ”๋”ฉํ•˜์—ฌ ์‚ฌ๋žŒ ์ด๋ฏธ์ง€๋ฅผ ์ƒ์„ฑํ•˜๋Š” ๋ฐฉ๋ฒ•์„ ์†Œ๊ฐœํ•ฉ๋‹ˆ๋‹ค. ๋…ผ๋ฌธ https://openaccess.thecvf.com/contentcvpr2018/papers/MaD

2021๋…„ 6์›” 16์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€

์ž๋ฃŒ๊ตฌ์กฐ์™€ ์•Œ๊ณ ๋ฆฌ์ฆ˜ ์†Œ๊ฐœ ๋ฐ Array

์ž๋ฃŒ๊ตฌ์กฐ์™€ ์•Œ๊ณ ๋ฆฌ์ฆ˜์„ ์†Œ๊ฐœํ•˜๊ณ  ๋ฐฐ์—ด์˜ 4๊ฐ€์ง€ ์—ฐ์‚ฐ์— ๋Œ€ํ•ด ์„ค๋ช…ํ•ฉ๋‹ˆ๋‹ค.

2021๋…„ 6์›” 15์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€
post-thumbnail

Pose-Guided Person Image Synthesis in the Non-Iconic Views

๋ณธ ๋…ผ๋ฌธ์€ ๊ธฐ์กด์˜ Person Generation์ด Iconicํ•œ View์—์„œ๋งŒ ์„ฑ๋Šฅ์ด ์ข‹์€ ๊ฒƒ์„ ๋ณด๊ณ , Non Iconicํ•œ ์ƒํ™ฉ์˜ ์ด๋ฏธ์ง€์—์„œ๋„ ์„ฑ๋Šฅ์„ ๋†’์ด๊ธฐ ์œ„ํ•ด ์ œ์•ˆ๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

2021๋…„ 6์›” 14์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€
post-thumbnail

Deformable GANs for Pose-based Human Image Generation

๋ณธ ๋…ผ๋ฌธ์€ ์‚ฌ๋žŒ ์ด๋ฏธ์ง€์—์„œ์˜ ํฌ์ฆˆ๋ฅผ ๋ฐ”๊ฟ€ ๋•Œ, ํฌ์ฆˆ๋“ค ์‚ฌ์ด์—์„œ misalignment ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๋Š” ๋ฐฉ๋ฒ•์— ๋Œ€ํ•ด์„œ ๋‹ค๋ฃน๋‹ˆ๋‹ค.

2021๋…„ 6์›” 11์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€
post-thumbnail

๋”ฅ๋Ÿฌ๋‹ ๊ธฐ์ดˆ(2)

โ€ข ์ธ๊ณต์‹ ๊ฒฝ๋ง (Neural Network)โ€ข ๊นŠ์€ ์ธต (Deep Layer)โ€ข ๋น„์„ ํ˜•์„ฑ (Non-Linearity)โ€ข ์ตœ์ ํ™” (Optimization)โ€ข ๋”ฅ๋Ÿฌ๋‹ ๋ชจ๋ธ์˜ ๊ณต๊ฐ„์  ์˜๋ฏธโ€ข ์ธ๊ณต์‹ ๊ฒฝ๋ง ๋ฌธ์ œํ•ด๊ฒฐ ํ”„๋กœ์„ธ์Šคโ€‚โ€‚ \- ๋ฌธ์ œ ์ •์˜ โ–ถ ๋ชจ๋ธ ๊ตฌ์„ฑ โ–ถ ๋ชจ๋ธ ํ•™์Šต โ–ถ ๋ฌธ์ œ

2021๋…„ 6์›” 10์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€
post-thumbnail

๋จธ์‹ ๋Ÿฌ๋‹ ๊ฐœ์š”

โ€ข ํ•™์Šต, ๋ฌธ์ œ ํ•ด๊ฒฐ, ํŒจํ„ด ์ธ์‹ ๋“ฑ ์ธ์ง€ ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๋Š” ์ปดํ“จํ„ฐ ๊ณตํ•™ ๋ถ„์•ผ.โ€ข 1943๋…„ ์ฒ˜์Œ์œผ๋กœ ๊ฐœ๋…์ด ๋“ฑ์žฅ.โ€ข ์ตœ๊ทผ ๋น…๋ฐ์ดํ„ฐ์™€ ์ปดํ“จํŒ… ์„ฑ๋Šฅ ํ–ฅ์ƒ์œผ๋กœ ์ธํ•ด ํญ๋ฐœ์ ์ธ ์„ฑ์žฅ ์ค‘.โ€ข Narrow AI โ€ƒโ†’์ข์€ ์˜์—ญ์—์„œ ํŠน์ •ํ•œ task๋งŒ ๊ฐ€๋Šฅํ•˜์ง€๋งŒ ์ •ํ™•๋„๋‚˜ ์†๋„๋Š” ์ธ๊ฐ„์„

2021๋…„ 6์›” 10์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€
post-thumbnail

Deep Photo Enhancer: Unpaired Learning for Image Enhancement from Photographs with GANs

์ด ๋…ผ๋ฌธ์€ ์ด๋ฏธ์ง€ ํ–ฅ์ƒ์„ ์œ„ํ•œ Unpaired learning ํ•™์Šต ๋ฐฉ๋ฒ•์„ ์ œ์•ˆํ•œ ๋…ผ๋ฌธ์ž…๋‹ˆ๋‹ค.

2021๋…„ 6์›” 10์ผ
ยท
0๊ฐœ์˜ ๋Œ“๊ธ€