K-Ruby-22-try-using-GAN

0b2426ff15397042caa863cc7965c091?s=47 Yuta Kurotaki
March 25, 2019
Research
2
170

 K-Ruby-22-try-using-GAN

2019/03/25 K-Ruby#22
GMOペパボ鹿児島オフィス
自由研究: GANを活用した新しいアイテム作成体験を考えていく

0b2426ff15397042caa863cc7965c091?s=128

Yuta Kurotaki

March 25, 2019
Tweet

More Decks by Yuta Kurotaki

See All by Yuta Kurotaki
0b2426ff15397042caa863cc7965c091?s=48 kurotaky
1
390
0b2426ff15397042caa863cc7965c091?s=48 kurotaky
0
21
0b2426ff15397042caa863cc7965c091?s=48 kurotaky
2
590
0b2426ff15397042caa863cc7965c091?s=48 kurotaky
1
4.5k
0b2426ff15397042caa863cc7965c091?s=48 kurotaky
0
48
0b2426ff15397042caa863cc7965c091?s=48 kurotaky
2
190
0b2426ff15397042caa863cc7965c091?s=48 kurotaky
0
59
0b2426ff15397042caa863cc7965c091?s=48 kurotaky
1
290
0b2426ff15397042caa863cc7965c091?s=48 kurotaky
1
390

Other Decks in Research

See All in Research
A42dd3541cd40296dcd8a5e6b4a01bef?s=48 scatterlab
0
970
99e9e6d2de62c373990ac1bd7c4defc5?s=48 hakubishin3
1
540
E2dd989b2ba0f83d8a981b9cb3197bf1?s=48 j20232
0
360
6325a9b34da54d5cbddb814c3987a2fe?s=48 chokkan
1
3.3k
874ff503a00697a857e198a0ebb8f55f?s=48 ailaboocu
0
130
847a328633b1df6b11cc2f72430025e6?s=48 ks91
0
100
A42dd3541cd40296dcd8a5e6b4a01bef?s=48 scatterlab
0
960
090f9e164337989b54dabd7fedfd39b4?s=48 flekschas
0
310
C29f097d23f8904532ca088ac23ce801?s=48 kayceesrk
0
1k
608638a6cfefbf0c2294a30e55634f6d?s=48 mihozono
15
8.6k
6e166ed0b8cf24a9d5aef105dacf5db7?s=48 usaito
4
1.4k
Bb6c3fc8c577710c72d03aeb4fa56bf6?s=48 masakat0
0
190

Featured

See All Featured
651dcfe41723207fbb0aa044a4f7c07f?s=48 dotmariusz
94
5.1k
3d4519fd34b76fb265fc0237f3792bd4?s=48 danielanewman
199
19k
3d4519fd34b76fb265fc0237f3792bd4?s=48 danielanewman
1
400
8081b26e05bb4354f7d65ffc34cbbd67?s=48 chriscoyier
498
130k
B6a8f005f39d23ffc930508ac9da68b9?s=48 vanstee
116
4.8k
C0b42577cfc2b321be3474618107e933?s=48 samanthasiow
55
6.1k
C86443373fbfe92996aa882d0d7a59f8?s=48 imathis
479
150k
20bfe76b3d6105641f879fe45cfc9272?s=48 bkeepers
408
57k
32de0bd2ba869609d26fd052a4622778?s=48 cromwellryan
101
5.8k
E13c31390e0369fcd5972292ce0e7b92?s=48 jnunemaker
PRO
40
4.4k
78b475797a14c84799063c7cd073962f?s=48 holman
288
130k
761be20b5ebd271008bcee1244fc5b52?s=48 matthewcrist
72
7.2k

Transcript

  1. ("/Λ׆༻ͨ͠৽͍͠ΞΠςϜ࡞੒ମݧΛߟ͍͑ͯ͘ ࠇ୍༔ଠ :VUB,VSPUBLJ  ,3VCZ ࣗ༝ݚڀ

  2. )VNBO$PNQVUFS*OUFSBDUJPO3FTFBSDIFS ࠇ୍༔ଠ!LVSPUBLZ 46;63*ࣄۀ෦ςΫχΧϧϦʔυ

  3. ࣗݾ঺հ ೥݄ΑΓ(.0ϖύϘגࣜձࣾʹͯιϑτ΢ΣΞΤϯδχΞͱͯ͠ෳ਺ͷ8FCαʔ Ϗεͷ։ൃΛ୲౰ɺݱࡏ͸ςΫχΧϧϦʔυͱͯ͠46;63*ͱ46;63*1FPQMFͷ։ൃ ͓ΑͼΤϯδχΞϦϯά૊৫ͷϚωʔδϝϯτΛߦ͍ͬͯΔɻ ϖύϘʹೖࣾͯ͠3VCZΛ࢖͍͸͡Ίͯϓϩάϥϛϯάָ͍ͬͯ͠ࢥ͖ͬͯͨਓͰ͢ʂ  5XJUUFS!LVSPUBLZ  'BDFCPPLLVSPUBLZ 

    (JU)VCLVSPUBLZ

  4. ٕज़ॻయ .4UBDLೖ໳ʹؔ͢ΔຊΛ ϖύϘ.BLF෦ͷΈΜͳͰॻ͖·ͨ͠

  5. 46;63* IUUQTTV[VSJKQ

  6. None

  7. ࣛࣇౡΦϑΟε01&/

  8. ΤϯδχΞ໊͕ࣛࣇౡΦϑΟεʹདྷ·͢ w݄೔͔ΒࣛࣇౡΦϑΟεʹདྷ·͢ʂ w,3VCZʹࢀՃ͢Δ༧ఆͳͷͰΑΖ͓͘͠Ͷ͕͍͠·͢ʂ w!IBZBQJͱ!LFOLFO ݄຤·Ͱ౦ژͰݚम͍ͯ͠·͢

  9. ࣗ༝ݚڀ

  10. ࣗ༝ݚڀ w)$*3FTFBSDIFSͱͯ͠ͷ׆ಈΛ։࢝ wਓͱίϯϐϡʔλݱ࣮ੈքͱͷؔ܎ʹ͍ͭͯߟ͑Δͧʂ wݚڀ։ൃͬΆ͍͜ͱΛগͣͭ͠΍͍ͬͯ͜͏ wlը૾ΛຕΞοϓ͢Δ͚ͩzΛٙͬͯΈΔ

  11. ਓ͕ؒը૾ΛΞοϓϩʔυ͠ͳ͍Ͱ ঎඼Λ࡞Δʹ͸ʁ

  12. None

  13. ը૾Λࣗಈੜ੒ͯ͠ΈΔ

  14. Πϝʔδ 46;63*"1*ܦ༝ͰΞΠςϜ࡞੒  l೜ऀεϦεϦ܅zͱ͍͏Ωʔϫʔυೖྗ l೜ऀεϦεϦ܅zͰը૾ݕࡧ ݕࡧ݁Ռͷը૾Λ5SBJOJOH4FUʹͯ͠ ("/Ͱը૾ੜ੒ l೜ऀεϦεϦ܅z ը૾ΛOຕऔಘɺֶश༻ʹDPOWFSU͢Δ

  15. ("/ (FOFSBUJWF"EWFSTBSJBM/FUXPSLT wఢରతੜ੒ωοτϫʔΫ wࠓճ͸%$("/ %FFQ$POWPMVUJPOBM("/ Λ࢖༻͢Δ *NBHFIUUQTNFEJVNDPNNMFWFSZUIJOHHFOFSBUJOHMFUUFSTVTJOHHFOFSBUJWFBEWFSTBSJBMOFUXPSLTHBOTCCFD %$("/("/ʹରͯ͠৞ΈࠐΈχϡʔϥϧωοτϫʔΫΛద༻͠ɺ͏·ֶ͘श͕੒ཱ͢ΔΑ͏ͳ΋ͷ

  16. ࣮ݧ w(PPHMFը૾ݕࡧͰ೜ऀεϦεϦ܅ͷը૾Λμ΢ϯϩʔυͯ͘͠Δ wຕͷը૾Λ5SBJOJOH4FUͰ༩͑ɺ("/Λ࢖ͬͯը૾Λੜ੒͢Δ w(PPHMF$PMBCPSBUPSZ͔Β(16Λ࢖ͬͯ%$("/Λ࣮ߦͨ͠ ࣮૷IUUQTHJUIVCDPNDBSQFEN%$("/UFOTPSqPX

  17. ݁Ռ wճ਺ΛॏͶΔͱεϦεϦ܅Ά͍ը૾͕ੜ੒͞Ε͖ͯͨ FQPDI FQPDI FQPDI

  18. None

  19. ·ͱΊ wϢʔβʔ͕ը૾ͷΞοϓϩʔυΛͤͣʹΞΠςϜΛ࡞੒͢Δख๏Λߟ͑ͨ w%$("/Λ࢖ͬͯը૾ੜ੒·ͰͰ͖ͨ w(16Λ࢖ͬͯ΍ͬͯ΋ɺͦΕͳΓʹ͕͔͔࣌ؒΔ wYͷը૾ຕΛFQPDIͰ࣌ؒ͘Β͍͔͔ͬͨ w(PPHMF$PMBCศར

  20. ࠓޙ wςΩετೖྗͰ͖Δ6*͔Βը૾ੜ੒ˠΞΠςϜ࡞੒·Ͱߦ͏ wࠓճ͸ίϯιʔϧ͔ΒεΫϦϓτͰୟ͍ͯΞΠςϜੜ੒ wύϥϝʔλΛͲ͏޻෉͢Ε͹ྑ͍ײ͡ʹͳΔͷ͔ࢼ͢ wࣗ਎ͷ਺ֶͷֶशΛܧଓ