Hcmute thiết kế hệ thống chuyển đổi giọng nói sang ngôn ngữ cử chỉ ứng dụng cho người khiếm thính

BỘ GIÁO DỤC VÀ ĐÀO TẠO TRƯỜNG ĐẠI HỌC SƯ PHẠM KỸ THUẬT THÀNH PHỐ HỒ CHÍ MINH CƠNG TRÌNH NCKH CẤP TRƯỜNG TRỌNG ĐIỂM THIẾT KẾ HỆ THỐNG CHUYỂN ĐỔI GIỌNG NĨI SANG NGƠN NGỮ CỬ CHỈ ỨNG DỤNG CHO NGƯỜI KHIẾM THÍNH S K C 0 9 MÃ SỐ: T2020-39TĐ S KC 0 Tp Hồ Chí Minh, tháng 4/2021 Luan van BỘ GIÁO DỤC VÀ ĐÀO TẠO TRƯỜNG ĐẠI HỌC SƯ PHẠM KỸ THUẬT THÀNH PHỐ HỒ CHÍ MINH BÁO CÁO TỔNG KẾT ĐỀ TÀI KH&CN CẤP TRƯỜNG TRỌNG ĐIỂM THIẾT KẾ HỆ THỐNG CHUYỂN ĐỔI GIỌNG NÓI SANG NGƠN NGỮ CỬ CHỈ ỨNG DỤNG CHO NGƯỜI KHIẾM THÍNH Mã số: T2020-39TĐ Chủ nhiệm đề tài: TS Trương Ngọc Sơn TP HCM, 04/2021 Luan van TRƯỜNG ĐẠI HỌC SƯ PHẠM KỸ THUẬT THÀNH PHỐ HỒ CHÍ MINH KHOA ĐIỆN – ĐIỆN TỬ BÁO CÁO TỔNG KẾT ĐỀ TÀI KH&CN CẤP TRƯỜNG TRỌNG ĐIỂM THIẾT KẾ HỆ THỐNG CHUYỂN ĐỔI GIỌNG NĨI SANG NGƠN NGỮ CỬ CHỈ ỨNG DỤNG CHO NGƯỜI KHIẾM THÍNH Mã số: T2020-39TĐ Chủ nhiệm đề tài: TS Trương Ngọc Sơn Thành viên đề tài: ThS Lê Minh Thành ThS Lê Minh TP HCM, 04/2021 Luan van DANH SÁCH THÀNH VIÊN THAM GIA ĐỀ TÀI Số TT Họ Tên Nhiệm vụ Trương Ngọc Sơn Chủ nhiệm Lê Minh Thành Thành viên Lê Minh Thành viên Luan van MỤC LỤC DANH MỤC BẢNG BIỂU DANH MỤC CÁC CHỮ VIẾT TẮT MỞ ĐẦU 1 Tổng quan Tính cấp thiết đề tài Mục tiêu đề tài Đối tượng, phạm vi nghiên cứu Phương pháp nghiên cứu Nội dung nghiên cứu Chương NHẬN DẠNG GIỌNG NÓI 1.1 Giói thiệu 1.2 Trích rút đặc trưng tín hiệu lời nói 1.3 Các mơ hình nhận dạng giọng nói 1.4 Mô hình âm học (acoustic model) 1.5 Mơ hình ngơn ngữ (language model) 12 1.6 Các mơ hình nhận dạng giọng nói 15 1.6.1 SPHINX 15 1.6.2 POCKETSPHINX 17 1.6.3 Mơ hình mạng nơ-ron học sâu - DeepSpeech 20 1.6.4 Mạng nơ-ron học sâu - ConvNet 22 1.6.5 Dịch vụ nhận dạng giọng nói Google (Google speech regnition) 24 1.6.6 Nhận xét 24 1.7 Giới thiệu ngôn ngữ cử 25 Chương 28 THIẾT KẾ HỆ THỐNG CHUYỂN ĐỔI GIỌNG NÓI SANG NGÔN NGỮ CỬ CHỈ 28 2.1 Thiết kế phần cứng 28 2.2 Thiết kế phần mềm xử lý 30 Chương 33 KẾT QUẢ NGHIÊN CỨU VÀ ỨNG DỤNG 33 3.1 Kết thực mơ hình 33 3.2 Đánh giá tốc độ đáp ứng hệ thống 34 3.3 Đánh giá độ xác hệ thống 35 Chương 36 KẾT LUẬN VÀ KIẾN NGHỊ 36 Luan van 5.1 Kết nghiên cứu 36 5.2 Kiến nghị định hướng nghiên cứu 36 TÀI LIỆU THAM KHẢO 37 PHỤ LỤC Bài báo thuộc danh mục sản phẩm đề tài Luan van DANH MỤC BẢNG BIỂU Bảng 1.1 Tập liệu huấn luyện 17 Bảng 1.2 Mơ hình âm học cho hệ thống nhận dạng tiếng nói Pocketsphinx 18 Bảng 1.3 Mơ hình ngơn ngữ 19 Luan van DANH MỤC CÁC CHỮ VIẾT TẮT CNN Convolutional Neural Network HMM Hidden Markov Model EM Expectation Maximization GMM Gaussian Mixture Mode MFCC Mel Frequency Cepstral Coefficents LPC Linear Prediction Cepstral API Application Programming Interface RNN Recurrent Neural Network Luan van TRƯỜNG ĐẠI HỌC SƯ PHẠM KỸ THUẬT CỘNG HOÀ XÃ HỘI CHỦ NGHĨA VIỆT NAM THÀNH PHỐ HỒ CHÍ MINH Độc lập - Tự - Hạnh phúc KHOA ĐIỆN – ĐIỆN TỬ Tp HCM, ngày 10 tháng 04 năm 2021 THƠNG TIN KẾT QUẢ NGHIÊN CỨU Thơng tin chung: - Tên đề tài: Thiết kế hệ thống chuyển đổi giọng nói sang ngơn ngữ cử ứng dụng cho người khiếm thính - Mã số: T2020-39TĐ - Chủ nhiệm: TS Trương Ngọc Sơn - Cơ quan chủ trì: Trường Đại học Sư phạm Kỹ thuật TP.HCM - Thời gian thực hiện: 12 tháng Mục tiêu: Thiết kế hệ thống nhận dạng giọng nói (tiếng Việt) chuyển sang ngôn ngữ cử dạng ảnh Tính sáng tạo: Thiết kế hệ thống chuyển đổi giọng nói sang ngơn ngữ có khả thực thi phần cứng có cấu hình thấp Kết nghiên cứu: Thiết kế hệ thống chuyển đổi giọng nói sang ngơn ngữ Hệ thống có tính linh hoạt, nhỏ gọn, tiêu thụ công suất thấp, hoạt động liên tục Thông tin chi tiết sản phẩm: a Sản phẩm khoa học: + Báo cáo khoa học (ghi rõ số lượng, giá trị khoa học): 01 báo cáo khoa học + Bài báo khoa học (ghi rõ đầy đủ tên tác giả, tên báo, tên tạp chí, số xuất bản, năm xuất bản): 01 báo chấp nhận đăng Tạp chí International Journal of Computer Science and Network Security, vol 21 no.3, March 2021 (ESCI) Luan van b Sản phẩm ứng dụng (bao gồm vẽ, mơ hình, thiết bị máy móc, phần mềm…, ghi rõ số lượng, quy cách, công suất….): Hiệu quả, phương thức chuyển giao kết nghiên cứu khả áp dụng: Trưởng Đơn vị Chủ nhiệm đề tài (ký, họ tên) (ký, họ tên) Luan van Luan van Luan van Luan van Luan van Luan van Luan van Luan van Luan van Luan van Luan van IJCSNS International Journal of Computer Science and Network Security, VOL.21 No.3, March 2021 37 A Low-Cost Speech to Sign Language Converter Minh Le, Thanh Minh Le, Vu Duc Bui, Son Ngoc Truong HCMC University of Technology and Education, Ho Chi Minh City, Vietnam Summary This paper presents a design of a speech to sign language converter for deaf and hard of hearing people The device is lowcost, low-power consumption, and it can be able to work entirely offline The speech recognition is implemented using an opensource API, Pocketsphinx library In this work, we proposed a context-oriented language model, which measures the similarity between the recognized speech and the predefined speech to decide the output The output speech is selected from the recommended speech stored in the database, which is the best match to the recognized speech The proposed context-oriented language model can improve the speech recognition rate by 21% for working entirely offline A decision module based on determining the similarity between the two texts using Levenshtein distance decides the output sign language The output sign language corresponding to the recognized speech is generated as a set of sequential images The speech to sign language converter is deployed on a Raspberry Pi Zero board for low-cost deaf assistive devices receive information from the real world In this paper, we experimentally present a design of a low-cost one-way portable translator, which can translate speech to sign language for deaf people The speech recognition is performed using the open-source library, Pocketsphinx, which can work entirely offline [10]-[13] Pocketsphinx is a small reconfigurable model which can be deployed on low-cost embedded systems for a mobile device In this experimental demonstration, we deploy the speech recognition model and the speech to sign language on a low-cost Raspberry Pi Zero [14]-[16] The optimized Pocketsphinx model has low accuracy because it employs the limited acoustic and language model To improve the accuracy, a context-oriented language model is proposed The proposed context-oriented language model is based on the Levenshtein Distance to measure the similarity between the recognized speech and the recommended speech Key words: Speech recognition, speech-to-text, sign language Design a speech to sign language converter Introduction Speech is a common way that humans use to communicate with each other However, it is not used for people who suffer from speech and hearing disabilities There are many people around the globe having disabilities in hearing and speaking There is an existing language used for such people, which is called sign language Sign language is a fully visual language with its grammar used for deaf and mute people [1] In sign language, hand gestures, head, body movements, and facial expressions are used by humans to convey the information [2] It is difficult for deaf people to understand the information from speech coming from normal people or even from media devices Therefore, a translator is necessary for minimizing the communication gap between hearing impaired and normal people The translator owns a speech recognition that translates speech to text and a sign language generator that converts text to appropriate sign language [3]-[5] Speech recognition is a complicated task The leading-edge speech recognition model is implemented using machine learning and deep neural network [6]-[9] However, deep neural networks are based on a massive number of computational tasks which consume a huge amount of power and processing time For this reason, a simple, portable, and low-cost translator is necessary for helping deaf people to A speech to sign language converter for deaf people must satisfy several requirements such as mobility, low power consumption, low cost, and high reliability There are three necessary modules inside such the device including speech-to-text module, language understanding module, and text to sign language converter module For low-cost devices, we utilize the Raspberry Pi Zero for the control unit The speech to text, language understanding, and text to sign language are deployed on the Raspberry Pi Zero board Fig The block diagram of the speech to sign language converter Manuscript received March 5, 2021 Manuscript revised March 20, 2021 https://doi.org/10.22937/IJCSNS.2021.21.3.5 Luan van 38 IJCSNS International Journal of Computer Science and Network Security, VOL.21 No.3, March 2021 Fig shows a block diagram of a speech to sign language converter The Raspberry Pi Zero is a low-cost embedded system being suitable for portable devices Raspberry Pi Zero has only one channel for output audio In this design, the input speech signal recorded from the microphone is passed to the Raspberry Pi Zero using a USB Audio Adapter, as shown in Fig Sign language is the animation composed of a set of sequential images being displayed on an LCD The system is powered by a 4000 mAh battery The device can work continuously hours Speech to text is an important task that decides the reliability of the device Recently, deep learning-based speech-to-text modules have been demonstrated as the leading edge model for automatic speech recognition systems However, the deep neural network is a resourcehungry platform since it is based on huge computational tasks For being suitable to be deployed on a low-cost computer such as Raspberry Pi Zero, we used an offline open-source speech-to-text module, Pocketsphinx instead We propose a context-oriented language model to improve speech recognition accuracy The Pocketsphinx is followed by the proposed context-oriented language model is shown in Fig Natural language processing is commonly used for language understanding tasks However, the natural language proposed is a complicated task that requires a huge resource In this design, we use a simple method that compares the recognized speech and the one stored in the database to decide the output Levenshtein Distance method is suitable for such a task The last task is the text to sign language converter Having decided the output, a sign language is generated as a set of consecutive images playing on the output device Fig shows the low-cost speech to sign language converter with proposed context-oriented language model Fig The proposed speech to sign language converter with context-oriented language model In Fig 2, the speech is recorded from the microphone and then enters the Raspberry Pi Zero where speech is converted to text by Pocketsphinx module Pocketsphinx is an optimized Sphinx for low-cost computers The proposed context-oriented language model is based on the Levenshtein Distance to measure the similarity of recognized speech and desired speech stored in the database The output of the context-oriented language model is the speech obtained from the recommended speech which is the best match to the recognized speech By using the proposed context-oriented language model, the recognized speech is corrected according to the expected speech stored in the database The corrected text is then entered into the language understanding module where the output sign language is decided Here the text is compared with the predefined text to determine which sign language output will be Experimental Results The proposed architecture with three modules of speech to text, language understanding, and text to sign language is deployed on a Raspberry Pi Zeros board for a low-cost speech to sign language converter The speech recognition accuracy is improved by using the proposed contextoriented language model which corrects the recognized speech In table 1, we demonstrate the operation of the proposed context-oriented language model Table I The proposed context-oriented language model Pocketsphinx output Database Similarity score What you doing What are you why you go why you go I go working I have to walk now What are you doing What are you doing What are you doing Where you go I am working I have to work now 0.86 0.9 0.5 0.8 0.83 0.89 Contextoriented language model output What are you doing What are you doing Where you go Where you go I am working I have to work now In table I, we evaluate the performance of the proposed context-oriented language model in enhancing the accuracy of offline speech recognition The recognized speech is compared with the recommended speech stored in the database to decide the output speech if the similarity score is higher than 0.8 The Levenshtein Distance is utilized to measure the similarity of two strings By doing this, the text converted from speech resulted from Pocketsphinx is corrected In table I, the recognized from Pocketsphinx is “What you doing”, it is similarly the Luan van IJCSNS International Journal of Computer Science and Network Security, VOL.21 No.3, March 2021 recommended speech of “What are you doing”, then the corrected output is “What are you doing” Similarly, the recognized speech of “why you go” better matches with “where you go” rather than “What are you doing”, the output is “where you go” Using the proposed contextoriented model, the speech recognition rate is improved significantly The database is composed of possible sentences To evaluate the speech recognition module, we measure the accuracy for 500 sentences from speakers The speech is recorded from the microphone in realtime The Pocketsphinx without the proposed context-oriented model can recognize the speech and convert it to text with an accuracy as high as 71% The Pocketsphins followed by the proposed context-oriented language model has accuracy as high as 92% The proposed context-oriented language model can improve the accuracy by 21% Having received the speech, the sign language is generated at the output Sign language is based on a set of sequential images as shown in Fig Fig The sign language of “what are you doing” is composed of a set of consecutive images Sign language corresponding to the text output from the speech to text module is a set of sequential images which specific meaning as shown in Fig The sign language is displayed on an LCD device Conclusion This paper presented a design of a speech to sign language converted for deaf people The device is mobility, low power consumption, and can work without an internet connection The speech recognition is implemented by using an open-source library, Pocketsphinx module To enhance the accuracy, we proposed a context-oriented language model, which measures the similarity between the recognized speech and the predefined speech to decide the output The proposed model can improve speech recognition accuracy by 21% A decision module is based on a similarity between the two texts using Levenshtein distance decides the output sign language Acknowledgments This work belongs to the project grant No: T2020-39TĐ, funded by Ho Chi Minh City University of Technology and Education, Vietnam 39 References [1] U Bellugi and S Fischer, “A comparison of sign language and spoken language” Cognition, vol 1, no 2–3, pp 173200, 1972 [2] O Aran and L Akarun, “Sign Language Processing and Interactive Tools for Sign Language Education,” 2007 IEEE 15th Signal Processing and Communications Applications, Eskisehir, 2007, pp 1-4 [3] L Boppana, R Ahamed, H Rane and R K Kodali, “Assistive Sign Language Converter for Deaf and Dumb,” 2019 International Conference on Internet of Things (iThings) and IEEE Green Computing and Communications (GreenCom) and IEEE Cyber, Physical and Social Computing (CPSCom) and IEEE Smart Data (SmartData), Atlanta, GA, USA, 2019, pp 302-307 [4] N C Camgoz, S Hadfield, O Koller, H Ney and R Bowden, “Neural Sign Language Translation,” 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, 2018, pp 7784-7793 [5] L Kau, W Su, P Yu and S Wei, “A real-time portable sign language translation system,” 2015 IEEE 58th International Midwest Symposium on Circuits and Systems (MWSCAS), Fort Collins, CO, 2015, pp 1-4 [6] P Lakkhanawannakun and C Noyunsan, "Speech Recognition using Deep Learning," 2019 34th International Technical Conference on Circuits/Systems, Computers and Communications (ITC-CSCC), JeJu, Korea (South), 2019, pp 1-4 [7] I Gavat and D Militaru, “Deep learning in acoustic modeling for Automatic Speech Recognition and Understanding - an overview,” 2015 International Conference on Speech Technology and Human-Computer Dialogue (SpeD), Bucharest, Romania, 2015, pp 1-8 [8] A Kumar, S Verma and H Mangla, “A Survey of Deep Learning Techniques in Speech Recognition,” 2018 International Conference on Advances in Computing, Communication Control and Networking (ICACCCN), Greater Noida, India, 2018, pp 179-185 [9] N K Mudaliar, K Hegde, A Ramesh and V Patil, “Visual Speech Recognition: A Deep Learning Approach,” 2020 5th International Conference on Communication and Electronics Systems (ICCES), Coimbatore, India, 2020, pp 1218-1221 [10] K Lee, H Hon, M Hwang, S Mahajan and R Reddy, “The SPHINX speech recognition system,” International Conference on Acoustics, Speech, and Signal Processing,, Glasgow, UK, 1989, pp 445-448 vol.1, doi: 10.1109/ICASSP.1989.266459 [11] K Lee, H Hon and R Reddy, “An overview of the SPHINX speech recognition system,” in IEEE Transactions on Acoustics, Speech, and Signal Processing, vol 38, no 1, pp 35-45, Jan 1990 [12] D Huggins-Daines, M Kumar, A Chan, A W Black, M Ravishankar and A I Rudnicky, “Pocketsphinx: A Free, Real-Time Continuous Speech Recognition System for Hand-Held Devices,” 2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings, Toulouse, 2006, pp I-I [13] B Lakdawala, F Khan, A Khan, Y Tomar, R Gupta and A Shaikh, “Voice to Text transcription using CMU Sphinx Luan van 40 IJCSNS International Journal of Computer Science and Network Security, VOL.21 No.3, March 2021 A mobile application for healthcare organization,” 2018 Second International Conference on Inventive Communication and Computational Technologies (ICICCT), Coimbatore, 2018, pp 749-753 [14] D B C Lima, R M B da Silva Lima, D de Farias Medeiros, R I S Pereira, C P de Souza and O Baiocchi, “A Performance Evaluation of Raspberry Pi Zero W Based Gateway Running MQTT Broker for IoT,” 2019 IEEE 10th Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON), Vancouver, BC, Canada, 2019, pp 0076-0081 [15] N S Yamanoor and S Yamanoor, “High quality, low cost education with the Raspberry Pi,” 2017 IEEE Global Humanitarian Technology Conference (GHTC), San Jose, CA, USA, 2017, pp 1-5 [16] A P Jadhav and V B Malode, “Raspberry PI Based OFFLINE MEDIA SERVER,” 2019 3rd International Conference on Computing Methodologies and Communication (ICCMC), Erode, India, 2019, pp 531-533 Luan van S K L 0 Luan van ... Luan van Chương THIẾT KẾ HỆ THỐNG CHUYỂN ĐỔI GIỌNG NĨI SANG NGƠN NGỮ CỬ CHỈ 2.1 Thiết kế phần cứng Hệ thống chuyển đổi giọng nói sang ngơn ngữ cử phục vụ cho người khiếm thính, hệ thống phải đạt... Hệ thống chuyển đổi giọng nói sang ngơn ngữ cử đáp ứng mục đích chuyển đổi từ giọng nói sang ngơn ngữ cử hỗ trợ cho người khiếm thính 5.2 Kiến nghị định hướng nghiên cứu - Kết nghiên cứu sử dụng. .. sau sử dụng mạng nơ ron nhân tạo để nhận dạng ngôn ngữ cử [4] Hệ thống chuyển đổi từ giọng nói sang ngơn ngữ cử hỗ trợ cho người khiếm thính thiết kế bao gồm hệ thống nhận dạng tiếng nói tự động,