Sciweavers

ICASSP
2011
IEEE
13 years 4 months ago
Exemplar-based Sparse Representation phone identification features
Exemplar-based techniques, such as k-nearest neighbors (kNNs) and Sparse Representations (SRs), can be used to model a test sample from a few training points in a dictionary set. ...
Tara N. Sainath, David Nahamoo, Bhuvana Ramabhadra...
ICASSP
2011
IEEE
13 years 4 months ago
Pronunciation variation modeling of non-native proper names by discriminative tree search
In this paper, the task of selecting the optimal subset of pronunciation variants from a set of automatically generated candidates is recast as a tree search problem. In this appr...
Line Adde, Torbjørn Svendsen
ICASSP
2011
IEEE
13 years 4 months ago
Hill climbing on speech lattices: A new rescoring framework
We describe a new approach for rescoring speech lattices — with long-span language models or wide-context acoustic models — that does not entail computationally intensive latt...
Ariya Rastrow, Markus Dreyer, Abhinav Sethy, Sanje...
ICASSP
2011
IEEE
13 years 4 months ago
A 3-layer coding scheme for biometry template protection based on spectral minutiae
—Spectral Minutiae (SM) representation enables the combination of minutiae-based fingerprint recognition systems with template protection schemes based on fuzzy commitment, but ...
Xiaoying Shao, Haiyun Xu, Raymond N. J. Veldhuis, ...
ICASSP
2011
IEEE
13 years 4 months ago
Accurate transcription of broadcast news speech using multiple noisy transcribers and unsupervised reliability metrics
Professional manual transcription of speech is an expensive and time consuming process. This paper focuses on the problem of combining noisy transcriptions from multiple non-exper...
Kartik Audhkhasi, Panayiotis G. Georgiou, Shrikant...
SPEECH
2010
89views more  SPEECH 2010»
13 years 7 months ago
Which words are hard to recognize? Prosodic, lexical, and disfluency factors that increase speech recognition error rates
Despite years of speech recognition research, little is known about which words tend to be misrecognized and why. Previous work has shown that errors increase for infrequent words...
Sharon Goldwater, Daniel Jurafsky, Christopher D. ...
EJASMP
2010
196views more  EJASMP 2010»
13 years 7 months ago
Optimizing Automatic Speech Recognition for Low-Proficient Non-Native Speakers
Computer Assisted Language Learning (CALL) applications for improving the oral skills of low-proficient learners have to cope with nonnative speech that is particularly challengin...
Joost van Doremalen, Catia Cucchiarini, Helmer Str...
ECCC
2010
153views more  ECCC 2010»
13 years 7 months ago
Local list decoding with a constant number of queries
Efremenko showed locally-decodable codes of sub-exponential length that can handle close to 1 6 fraction of errors. In this paper we show that the same codes can be locally unique...
Avraham Ben-Aroya, Klim Efremenko, Amnon Ta-Shma
ICDAR
2009
IEEE
13 years 10 months ago
Document Content Extraction Using Automatically Discovered Features
We report an automatic feature discovery method that achieves results comparable to a manually chosen, larger feature set on a document image content extraction problem: the locat...
Sui-Yu Wang, Henry S. Baird, Chang An
ASPDAC
2010
ACM
152views Hardware» more  ASPDAC 2010»
13 years 10 months ago
Slack redistribution for graceful degradation under voltage overscaling
Modern digital IC designs have a critical operating point, or "wall of slack", that limits voltage scaling. Even with an errortolerance mechanism, scaling voltage below a...
Andrew B. Kahng, Seokhyeong Kang, Rakesh Kumar, Jo...