ETRI-Knowledge Sharing Plaform

Acoustic Model Automatic Speech Recognition(ASR) Korean Speech Self-Improving Deep Neural Network(DNN) Error Rate Speech Recognition System Beam Search Artificial Intelligence System Signal Noise Ratio(SNR) Real-World Data Speech Database Data Augmentation Robust Speech Recognition Based System State-Of-The-Art Speech Features Transformer-Based Deep Learning(DL) Connectionist Temporal Classification(CTC) Noisy Environments Bidirectional Lstm Barrier-Free Nonlinear Manifold Audio Data Audiovisual Dataset Assessment Method Augmentation Techniques Average Latency Conversational Speech Recognition Car Navigation System Context-Sensitive Enhancement Techniques Endpoint Detection Algorithm Hierarchical Detection GMM-HMM Human-Robot Interface Low-Resource Low-Power Model Structure Objective Function Performance Improvement Parameter Reduction Performance Improved Viterbi Decoding High Accuracy Response Latency Raspberry Pi Real-Time Inference Truncated Singular Value Decomposition Training Model Automatic Translation Direct Detection Decision Fusion Error Reduction Embedded Device Gaussian Mixture Model(GMM) High Performance Localization Method Large Vocabulary Model Parameter Multiple Speakers Need For Real Environment User-Perceived Latency Server-Client Spectral Ratio Speech Recognition Accuracy Blind Source Separation(BSS) Cross-Modality End-To-End Speech Recognition Field Environment Linear Regression Limited Resources Long Sequence Part Of Speech(POS) Post-Processing Method Reduction Technique Real-Time Operation Room Acoustic Voice Activity Detector Real-World Applications Signal-To-Noise Audio Signal Augmentation Method Computing Technology Detection Accuracy Efficient Algorithms Factorization Approach Fine-Tuning Hidden Markov Model Korean Language Non-Stationary Noise Robot System Separation Device Recognition Method Speech Enhancement Search Task Baseline System Channel Selection Frequency Range Frame Skipping Localization Algorithm Mobile Application(APP) Matrix Factorization Manifold Structure Network-Based Optimized Model Overlapped Speech Detection Phone Recognition Proficiency Assessment Web Speech Art Performance Edge Devices Formant Frequencies Far-Field Low-Rank Matrix Low Latency Learning Methods Online Speech Recognition Sound Source Localization Recognition Model Spectral Envelope Self-Supervised Learning Automatic Speech Recognizer Acoustic Features Bidirectional Long Short-Term Memory Based Approach Current State Digital Signal Processing Euclidean Norm Learning-Based Reduction Rate Voice Search Transformer Model System Accuracy Target Matrix Audio And Video Confidence Measure Computational Complexity Continuous Speech Computation Method Cognitive Computing Iot Systems Log Data Language Barrier Mobile Speech Recognition Manifold Regularized Noise Separation Non-Native Speakers Overall Performance Time Factor Speech Feature Extraction Recognition Performance Short Message