PK PyTorchKR Papers
Papers with Code · CC BY-SA 4.0

Keyword Spotting

130 평가 기록 13 datasets

In speech processing, keyword spotting deals with the identification of keywords in utterances.

<span style="color:grey; opacity: 0.6">( Image credit: Simon Grest )</span>

개요는 각 항목의 일부입니다. 위 탭에서 전체 기록을 페이지별로 탐색할 수 있습니다.

벤치마크

Papers with Code 아카이브의 데이터셋별 기록 예시입니다. 평가 조건과 지표 방향이 검증되지 않아 순위를 의미하지 않습니다.

Dataset기록된 모델 예시MetricPaperCode
ELiRF Fusion+Length(All Queries)
Cnxe: 0.6023MinCnxe: 0.5977ATWV: 0.3792MTWV: 0.3801
논문 -
TripletLoss-res15
Google Speech Commands V1 12: 98.56Google Speech Commands V2 12: 98.37Google Speech Commands V2 35: 97.0
논문 코드
HEiMDaL
Error Rate: 0.45%
논문 -
FKD
Res26
Accuracy: 95.88
논문 코드
QuaternionNeuralNetwork
Accuracy (10-fold): 98.53
논문 코드
CP-ResNet(ch64) w/ SSN(S=2, A=Sub)
Accuracy: 83.6% ±0.07
논문 -
TensorFlow's model version 2
TFMA: 89.7
논문 코드
1D-ConvNet
Accuracy (%): 93.7
논문 -
Quaternion Neural Networks
Accuracy(10-fold): 98.53
논문 코드
MicroNet-KWS-L
Accuracy: 95.3Latency (STM32F746ZG): 0.610128
논문 코드
Transpotter
Top-1 Accuracy: 65Top-5 Accuracy: 87.1mAP: 69.2mAP IOU@0.5: 68.3
논문 코드
Transpotter
Top-1 Accuracy: 52Top-5 Accuracy: 77.1mAP: 55.4mAP IOU@0.5: 53.6
논문 코드

데이터셋

13개
QUESSTGoogle Speech Commandshey SiriFKDGoogle Speech Commands V2 35TAU Urban Acoustic Scenes 2019TensorFlowVoxForgeGoogle Speech Commands (v2)Google Speech Commands V2 12LRS2LRS3-TED

논문

라이브러리

Papers with Code 리더보드 row에 연결된 구현 저장소입니다.

github.com/google-research/google-research/tree/master/kws_streaming 3 논문
github.com/aau-es-ml/ssl_noise-robust_kws 1 논문
github.com/Arizona-Voice/Arizona-spotting 1 논문
github.com/ARM-software/keyword-transformer 1 논문
github.com/EscVM/EscVM_YT/blob/master/Notebooks/1%20-%20TF2.X%20DeepAI-Quickie/tf_2_keyword_transformer.ipynb 1 논문
github.com/holgerbovbjerg/data2vec-kws 1 논문
github.com/ID56/Torch-KWT 1 논문
github.com/intelligentmachines/keyword_spotting_transformer 1 논문

하위 태스크

VK Visual Keyword Spotting 3 데이터셋 · 3 rows SK Small-Footprint Keyword Spotting 0 데이터셋 · 0 rows