Jump to content

Nsụgharị igwe nke asụsụ ndị ogbi

Shí Wikipedia, njikotá édémédé nke onyobulạ
Nsụgharị igwe nke asụsụ ndị ogbi

Ntụgharị igwe nke asụsụ ndị ogbi emeela, ọ bụ ezie na ọ dị obere, kemgbe 1977. Mgbe ọrụ nyocha nwere ihe ịga nke ọma na-ejikọta mkpụrụedemede Bekee site na keyboard gaa na mkpụrụedemede ASL nke a na-eme ka ọ dị n'aka robot. Nkà na ụzụ ndị a na-asụgharị asụsụ ndị aka n'asụsụ e dere ede ma ọ bụ nke a na-ekwu, na asụsụ e dere ede ọ bụ nke e kwuru iji asụsụ ndị aka, na-enweghị onye ntụgharị mmadụ. Asụsụ ndị ogbi nwere ọdịdị ụdaume dị iche iche karịa asụsụ ndị a na-asụ, nke mepụtara ihe mgbochi maka ndị mmepe. Ndị mmepe na-eji ọhụụ kọmputa na mmụta igwe iji mata ụfọdụ akụkụ phonological na epentheses [1] pụrụ iche maka asụsụ ndị ogbi, na njirimara okwu na nhazi asụsụ okike na-enye ohere nkwurịta okwu mmekọrịta n'etiti ndị na-anụ ihe na ndị ntị chiri.

Nkà na ụzụ ntụgharị asụsụ ndị ogbi nwere oke n'otu ụzọ ahụ dịka ntụgharị asụsụ a na-asụ. Ọ dịghị onye nwere ike ịsụgharị ya n'ụzọ ziri ezi 100%. N'ezie, teknụzụ ntụgharị asụsụ ndị ogbi nọ n'azụ ndị ibe ha na-asụ. Nke a bụ, n'ụzọ na-enweghị isi, n'ihi eziokwu ahụ bụ na asụsụ ndị a na-ahụ anya nwere ọtụtụ articulators. Ebe a na-asụ asụsụ site na ụda olu, a na-ekwupụta asụsụ ndị aka site na aka, ogwe aka, isi, ubu, ahụ, na akụkụ ihu. Nkwupụta a nwere ọtụtụ ọwa na-eme ka ịsụgharị asụsụ ndị ogbi sie ike. Ihe ịma aka ọzọ maka asụsụ ndị ogbi MT bụ eziokwu ahụ bụ na enweghị usoro ederede maka asụsụ ndị aka. E nwere usoro akara mana ọ nweghị usoro ederede nke ndị ntị chiri mba ụwa nabatara n'ọtụtụ ebe, nke na enwere ike iwere ya dị ka 'ụdị ederede' nke asụsụ ndị ogbi enyere. A na-edekọ asụsụ ndị ogbi n'ụdị vidiyo dị iche iche. Enweghị akara ọla edo nke buru ibu maka SMT, dịka ọmụmaatụ.

Akụkọ ihe mere eme

[dezie | dezie ebe o si]

Akụkọ ihe mere eme nke ntụgharị asụsụ ndị ogbi na-akpaghị aka malitere site na mmepe nke ngwaike dịka aka robot na-asụ mkpịsị aka. [2]'afọ 1977, ọrụ aka mkpịsị aka a na-akpọ RALPH (nke dị mkpirikpi maka "Robotic Alphabet") mepụtara aka robot nke nwere ike ịsụgharị mkpụrụedemede n'ime mkpịsị ụkwụ. [3] oge na-aga, iji gloves na sensọ mmegharị ghọrọ ihe a ma ama, a mụkwara ụfọdụ ọrụ dịka CyberGlove na VPL Data Glove. Ngwaọrụ a na-eyi na-eme ka o kwe omume ijide ọdịdị aka na mmegharị aka nke ndị bịanyere aka site n'enyemaka nke ngwanrọ kọmputa. Otú [3] dị, site na mmepe nke ọhụụ kọmputa, a na-eji igwefoto dochie ngwaọrụ ndị a na-eyi n'ihi arụmọrụ ha na obere ihe mgbochi anụ ahụ na ndị na-etinye aka. Iji dozie data a[4] site na ngwaọrụ, ndị na-eme nchọpụta mejupụtara netwọkụ akwara dịka Stuttgart Neural Network Simulator [1] maka njirimara njirimara na ọrụ dịka CyberGlove. Ndị na-eme nchọpụta na-ejikwa ọtụtụ ụzọ ndị ọzọ maka ịmata akara. [5] ọmụmaatụ, a na-eji Hidden Markov Models nyochaa data na ọnụ ọgụgụ, [1] na GRASP na mmemme mmụta igwe ndị ọzọ na-eji usoro ọzụzụ iji melite izi ezi nke njirimara akara. [3] Njikọ nke teknụzụ ndị [6] na-apụghị iyi ejiji dị ka igwefoto na ndị na-achịkwa Leap Motion egosila na ha na-abawanye ikike nke njirimara asụsụ ndị ogbi na ngwanrọ ntụgharị.

Ihe ịrịba ama n'olu dara ụda

[dezie | dezie ebe o si]

SignAloud bụ teknụzụ nke na-agụnye gloves abụọ nke otu ụmụ akwụkwọ na Mahadum Washington mere nke na-asụgharị [7] American Sign Language (ASL) n'asụsụ Bekee. [8] Na Febụwarị 2015 Thomas Pryor, nwa akwụkwọ na-anụ ihe na Mahadum Washington, mepụtara prototype mbụ maka ngwaọrụ a na Hack Arizona, hackathon na Mahaduma Arizona. Pryor gara n'ihu na-emepụta ihe ahụ ma n'ọnwa Ọktoba 2015, Pryor wetara Navid Azodi na ọrụ SignAloud maka ahịa na enyemaka na mmekọrịta ọha na eze. Azodi nwere ọgaranya [9] itinye aka na nchịkwa azụmahịa, ebe Pryor nwere ọgarịa nke ahụmịhe na injinia. [10]'ọnwa Mee afọ 2016, ha abụọ gwara NPR na ha na ndị na-eji ASL arụkọ ọrụ nke ọma ka ha wee nwee ike ịghọta ndị na-ege ha ntị nke ọma ma hazie ngwaahịa ha maka mkpa nke ndị a karịa mkpa a na-eche. Otú ọ dị, ọ dịghị nsụgharị ọzọ a tọhapụrụ kemgbe ahụ. Ihe nchoputa a bu otu n'ime asaa meriri Lemelson-MIT Student Prize, nke na-achọ inye onyinye na ịkụ aka na-eto eto. Ihe ha mepụtara dara n'okpuru ụdị "Jiri ya!" nke onyinye ahụ nke gụnyere ọganihu teknụzụ na ngwaahịa ndị dị ugbu a. [11][12] nyere ha $ 10,000. [1] [2]

Gloves nwere ihe mmetụta nke na-eso ndị ọrụ na-agagharị aka wee zipụ data na sistemụ kọmputa site na Bluetooth. Usoro kọmputa ahụ [10]-enyocha data ahụ ma mee ka ọ kwekọọ na okwu Bekee, nke olu dijitalụ na-ekwu n'olu dara ụda. Gloves enweghị ikike maka ntinye Bekee edere na mmepụta mmegharị glove ma ọ bụ ikike ịnụ asụsụ wee bịanye aka na ya na onye ntị chiri, nke pụtara na ha anaghị enye nkwurịta okwu. Ngwaọrụ ahụ a[13] etinye ihu ihu na akara ndị ọzọ na-abụghị nke aka nke asụsụ ndị ogbi, nke nwere ike ịgbanwe nkọwa n'ezie site na ASL.

Nkwupụta

[dezie | dezie ebe o si]

[14] (WebLibras) [15] bụ ngwanrọ kọmputa nke nwere ike ịsụgharị ma ederede na olu n'asụsụ Portuguese Libras (Asụsụ Ogbi Portuguese) "na ebumnuche nke imeziwanye nkwurịta okwu n'etiti ndị ntị chiri na ndị na-anụ ihe. " [1] E nwere mbipụta beta ugbu a na mmepụta maka Asụsụ Ogbi America. Ndị otu mbụ malitere ọrụ ahụ na 2010 site na njikọta nke ndị ọkachamara gụnyere ndị ọkà mmụta asụsụ, ndị na-emepụta ihe, ndị mmemme, na ndị nsụgharị, ma ndị na-anụ ihe ma ndị ntị chiri. Ndị otu ahụ malitere na Mahadum Federal nke Pernambuco (UFPE) site n'otu ìgwè ụmụ akwụkwọ na-etinye aka na ọrụ sayensị kọmputa. Ìgwè ahụ nwere onye otu ntị chiri nke na-esiri ndị ọzọ ike ikwurịta okwu. Iji mezue ọrụ ahụ [16] nyere onye otu ahụ aka ikwurịta okwu, otu ahụ mepụtara Proativa Soluções ma na-aga n'ihu kemgbe ahụ. Ụdị beta nke ugbu a na American Sign Language pere mpe. Dịka ọmụmaatụ, enwere ngalaba akwụkwọ ọkọwa okwu na naanị okwu dị n'okpuru mkpụrụedemede 'j' bụ 'jump'. Ọ bụrụ na ejighi okwu ahụ emepụta ngwaọrụ ahụ, mgbe ahụ avatar dijitalụ ga-ede okwu ahụ. Mmelite ikpeazụ nke ngwa ahụ [17] na June 2016, mana ProDeaf egosila n'ihe karịrị akụkọ 400 gafee ụlọ ọrụ mgbasa ozi kachasị ewu ewu na mba ahụ.

Ngwa ahụ enweghị ike ịgụ asụsụ ndị ogbi ma gbanwee ya ka ọ bụrụ okwu ma ọ bụ ederede, yabụ ọ na-eje ozi naanị dị ka nkwurịta okwu otu ụzọ. Tụkwasị na nke a, onye ọrụ ahụ enweghị ike ịbanye na ngwa ahụ ma nata nsụgharị Bekee n'ụdị ọ bụla, dịka Bekee ka dị na mbipụta beta.

Onye ntụgharị asụsụ ogbi nke Kinect

[dezie | dezie ebe o si]

Kemgbe afọ 2012, ndị na-eme nchọpụta sitere na Chinese Academy of Sciences na ndị ọkachamara na agụmakwụkwọ ndị ntị chiri site na Mahadum Beijing Union na China anọwo na-arụkọ ọrụ na ndị otu Microsoft Research Asian iji mepụta Kinect Sign Language Translator . [18] Onye ntụgharị nwere ụzọ abụọ: ụzọ onye ntụgharị na ụzọ nkwurịta okwu. Ọnọdụ onye ntụgharị nwere ike ịsụgharị otu okwu site na akara gaa na okwu edere ede na vice versa. Ọnọdụ nkwurịta okwu nwere ike ịsụgharị ahịrịokwu zuru ezu na mkparịta ụka ahụ nwere ike ịmegharị na-akpaghị aka site na iji avatar 3D. Ọnọdụ onye ntụgharị nwekwara ike ịchọpụta ọnọdụ na ọdịdị aka nke onye bịanyere aka yana usoro mmegharị site na iji teknụzụ nke mmụta igwe, njirimara njirimara, na Ọhụụ kọmputa. Ngwaọrụ ahụ [19]-enyekwa ohere maka Nkwupụta okwu n'ihi na teknụzụ njirimara okwu na-enye ohere ka a sụgharịa asụsụ a na-asụ n'asụsụ ndị ogbi na 3D modeling avatar nwere ike ịbịaghachi na ndị ntị chiri.

A malitere ọrụ mbụ ahụ na China dabere na ịsụgharị asụsụ ogbi Chinese. N'afọ 2013, e gosipụtara ọrụ ahụ na Microsoft Research Faculty Summit na nzukọ ụlọ ọrụ Microsoft. [20] [21] ọ dị ugbu a, ndị na-eme nchọpụta na United States na-arụkwa ọrụ a iji mezuo nsụgharị American Sign Language. Ka ọ dị ugbu a, ngwaọrụ ahụ ka bụ prototype, na izi ezi nke ntụgharị na ọnọdụ nkwurịta okwu ka ezughi oke.

Ihe ịrịba ama

[dezie | dezie ebe o si]

SignAll [22] bụ usoro ntụgharị asụsụ ndị ogbi na-akpaghị aka nke Dolphio Technologies [23] nyere na Hungary. Ndị otu ahụ na-ebute ụzọ na ntụgharị asụsụ ndị ogbi nke mbụ, dabere na ọhụụ kọmputa na nhazi asụsụ okike (NLP), iji mee ka nkwurịta okwu kwa ụbọchị n'etiti ndị na-anụ ihe na-eji Bekee na ndị ntị chiri ma ọ bụ ndị na-adịghị anụ ihe na-ejikarị ASL. " Usoro SignAll na-eji Kinect sitere na Microsoft na kamera weebụ ndị ọzọ jikọtara na ihe mmetụta miri emi na kọmputa. Nkà na ụzụ Ọhụụ kọmputa nwere ike ịmata ọdịdị aka na mmegharị nke onye na-etinye aka, usoro nhazi asụsụ sitere n'okike na-agbanwe data anakọtara site na ọhụụ kọmputa n'ime okwu Bekee dị mfe. Onye mepụtara ngwaọrụ ahụ bụ onye ntị chiri na ndị ọzọ na-arụ ọrụ ahụ nwere ọtụtụ ndị injinia na ndị ọkachamara asụsụ sitere na ndị ntị chiri na obodo ndị na-anụ ihe. Nkà na ụzụ ahụ nwere ikike itinye akụkụ ise niile nke ASL, nke na-enyere ngwaọrụ aka ịkọwa onye bịanyere aka n'ụzọ ziri ezi. Sig[24] akwadowo site na ọtụtụ ụlọ ọrụ gụnyere Deloitte na LT-innovate ma mepụta mmekọrịta na Microsoft Bizspark na Hungary's Renewal. [25] na-eji teknụzụ a eme ihe ugbu a na Fort Bend Christian Academy na Sugar Land, Texas na Sam Houston State University . [1]

MotionSavvy

[dezie | dezie ebe o si]

MotionSavvy [26] bụ asụsụ ndị ogbi mbụ na usoro olu. [27] mepụtara ngwaọrụ ahụ na 2012 site n'aka otu ìgwè sitere na Rochester Institute of Technology / National Technical Institute for the Deaf ma "site na Leap Motion accelerator AXLR8R". [1] Ndị otu ahụ jiri igbe mbadamba nke na-eji ike nke onye na-achịkwa Leap Moation. [28] bụ ụmụ akwụkwọ ntị chiri si n'ụlọ akwụkwọ agụmakwụkwọ ndị ntị chiri mepụtara ndị otu mmadụ isii ahụ. Ngwaọrụ ahụ bụ ugbu a otu n'ime naanị ngwaọrụ nkwurịta okwu abụọ naanị maka Asụsụ Ogbi nke America. Ọ na-enye ndị ntị chiri ohere ịbịanye aka na ngwaọrụ nke a na-akọwa ma ọ bụ vice versa, na-ewere Bekee a na-asụ ma na-akọwaa nke ahụ n'asụsụ ndị ogbi nke America. Ngwaọrụ ahụ na-ebugharị maka $ 198. Ụfọdụ atụmatụ ndị ọzọ gụnyere ikike imekọrịta ihe, nzaghachi oge ndụ, onye na-ewu akara, na akara igwe.

Oge ọ bụla enyochawo ngwaọrụ ahụ site na magazin teknụzụ ruo Time. Wired kwuru, "Ọ bụghị ihe siri ike ịhụ etu teknụzụ dị ka [UNI] nwere ike isi bụrụ" nakwa na "[UNI] kụrụ m dị ka ụdị anwansi. "Katy Steinmetz na TIME kwuru, "Thenkọnzụ a nwere ike ịgbanwe ụzọ ndị ntị chiri si ebi ndụ. " Sean Buckley na Engadget kwuru, "UNI nwere ike ịghọ Ngwá ọrụ nkwurịta okwu dị egwu".

Edensibia

[dezie | dezie ebe o si]
  1. Mocialov (2017). "Towards Continuous Sign Language Recognition with Deep Learning". Creating Meaning with Robot Assistants: The Gap Left by Smart Devices (IEEE-RAS International Conference on Humanoid Robots). 
  2. Jaffe (August 1994). "Evolution of mechanical fingerspelling hands for people who are deaf-blind.". Journal of Rehabilitation Research and Development 31 (3): 236–244. PMID 7965881. 
  3. 3.0 3.1 3.2 Parton (12 October 2005). "Sign Language Recognition and Translation: A Multidisciplined Approach From the Field of Artificial Intelligence". Journal of Deaf Studies and Deaf Education 11 (1): 94–101. DOI:10.1093/deafed/enj003. PMID 16192405. 
  4. Weissmann (1999). "Gesture recognition for virtual reality applications using data gloves and neural networks", IJCNN'99. International Joint Conference on Neural Networks. Proceedings (Cat. No.99CH36339), 2043–2046. DOI:10.1109/IJCNN.1999.832699. ISBN 978-0-7803-5529-3. 
  5. Bowden (2003). "Vision based Interpretation of Natural Sign Languages". 
  6. Bird (9 September 2020). "British Sign Language Recognition via Late Fusion of Computer Vision and Leap Motion with Transfer Learning to American Sign Language". Sensors 20 (18): 5151. DOI:10.3390/s20185151. PMID 32917024. 
  7. What is the difference between translation and transliteration. english.stackexchange.com. Retrieved on 2017-04-06.
  8. SignAloud. Archived from the original on 2020-09-21. Retrieved on 2024-02-24.
  9. Thomas Pryor and Navid Azodi | Lemelson-MIT Program. lemelson.mit.edu. Archived from the original on 2020-09-21. Retrieved on 2019-07-04.
  10. 10.0 10.1 "These Gloves Offer A Modern Twist On Sign Language", All Tech Considered, NPR, 17 May 2016.
  11. UW undergraduate team wins $10,000 Lemelson-MIT Student Prize for gloves that translate sign language. University of Washington (2016-04-12). Retrieved on 2017-04-09.
  12. Collegiate Inventors Awarded Lemelson-MIT Student Prize. Lemelson-MIT Program. Archived from the original on 2021-01-13. Retrieved on 2017-03-09.
  13. Nonmanual markers in American Sign Language (ASL). www.lifeprint.com. Retrieved on 2017-04-06.
  14. ProDeaf. www.prodeaf.net. Retrieved on 2017-03-09.
  15. ProDeaf. prodeaf.net. Retrieved on 2017-04-09.
  16. ProDeaf. www.prodeaf.net. Retrieved on 2017-03-16.
  17. ProDeaf Tradutor para Libras on the App Store (en). App Store. Retrieved on 2017-03-09.
  18. Chen. Kinect Sign Language Translator expands communication possibilities. Microsoft Research Connections. Archived from the original on 29 March 2014.
  19. Chai (2013). "Sign Language Recognition and Translation with Kinect". 
  20. Kinect Sign Language Translator. Microsoft (29 October 2013).
  21. Zafrulla (2011). "American sign language recognition with the kinect", Proceedings of the 13th international conference on multimodal interfaces - ICMI '11. DOI:10.1145/2070481.2070532. ISBN 978-1-4503-0641-6. 
  22. SignAll. We translate sign language. Automatically. (en). www.signall.us. Archived from the original on 2021-02-02. Retrieved on 2017-04-09.
  23. Dolphio | Unique IT Technologies (en-US). www.dolphio.hu. Retrieved on 2017-04-06.
  24. SignAll. We translate sign language. Automatically. (en). www.signall.us. Archived from the original on 2021-02-02. Retrieved on 2017-03-09.
  25. Fort Bend Christian Academy American Sign Language Program Pilots New Technology | Fort Bend Focus Magazine (en-US). Retrieved on 2023-08-08.
  26. MotionSavvy UNI: 1st sign language to voice system (en-US). Indiegogo. Retrieved on 2017-03-09.
  27. Rochester Institute of Technology (RIT) (en). Rochester Institute of Technology (RIT). Retrieved on 2017-04-06.
  28. Tsotsis (6 June 2014). MotionSavvy Is A Tablet App That Understands Sign Language. TechCrunch. Retrieved on 2017-04-09.