Papers by Timothy Israel Santos
2019 IEEE 4th International Conference on Big Data Analytics (ICBDA), 2019
The use of Deep Neural Network (DNN) models for Visual Speech Recognition (VSR) has recently been... more The use of Deep Neural Network (DNN) models for Visual Speech Recognition (VSR) has recently been gaining traction. The use of more complex DNN models have greatly increased accuracy performance but have the downside of very poor explainability. There is still much room for improvement in using DNN models for VSR in comparison to audio-only speech recognition. Being able to explain the model and its predictions would be beneficial for improving its performance, and the explainability of predictions are important in VSR in order to further improve the model design and handling of real-world data. This paper highlights various deep learning techniques for visual speech recognition and reports on experiments using feature visualisation techniques for these models, successfully demonstrating that CNNs are self-learning features consistent with what we would expect.
multilingualphilippines.com
The implementation of Mother Tongue-Based Multilingual
Education (MTBMLE) will require definitive... more The implementation of Mother Tongue-Based Multilingual
Education (MTBMLE) will require definitive rules for
orthography and grammar. While there are such rules for some
Philippine languages, there is a need to determine the agreement
and points of departure between the rules and the usage to avoid
confusion. This study makes an objective analysis of the levels of
agreement, in terms of grammar and orthographic rules, between
reference books and actual usage as evidenced from web-mined
text corpora for three major Philippine languages, namely
Filipino, Cebuano-Visayan and Ilokano. A list of language rules
on grammar and orthography were selected from standard
reference books for each of the aforementioned languages.
Alternative forms of usage for each selected language rule were
identified, and frequency counts were made, to be used as bases
for a comparative analysis between the rules prescribed by
standard reference books and actual language usage. The
techniques used in this study are important in language education,
serving to identify areas of confusion in language use in aspects of
grammar and orthography.
multilingualphilippines.com
Embedded in the human perception is the distinction of speech rhythm to discriminate between lang... more Embedded in the human perception is the distinction of speech rhythm to discriminate between languages. The rhythm of Filipino speech is given empirical validation in this paper using the computational and statistical methods prescribed by Ramus, Nespor, and Mehler (RNM) implemented on the Filipino Speech Corpus of the Digital Signal Processing Laboratory. Following the two sub-grouped convention, syllable-timed and stress-timed, the Filipino language was shown to be syllable-timed, and in the extended classification system, it was shown to be Mora-timed. The study was intended for the comprehensive modeling of the prosodic parameters of Filipino Speech for a natural-sounding Filipino Text-to-Speech (TTS) system, and a robust Automatic Speech Recognition system. The paper was already published and presented in a digital speech processing conference, but it was recently brought to our attention that the study has major implications to the linguistic description of the Filipino language, and to the recently ratified mother-tongue based multilingual education. Similar to digital speech systems, a person studying a second language (L2) will decide on the segment units he or she will use. If not guided properly, the student will be segmenting speech according to the rhythm of the native tongue (L1). Proper awareness of such speech parameters is very important for the student to avoid confusion, and especially for the teacher who will need to come up with an appropriate program to facilitate effective language acquisition.
Proceedings of the 25th Pacific Asia Conference on Language, Information and Computation. 2011., Jan 1, 2011
This study incorporates computational and perceptual methods to classify
Filipino speech rhythm. ... more This study incorporates computational and perceptual methods to classify
Filipino speech rhythm. Speech rhythm may be described as a language‟s distinguishing
durational sound pattern, resulting from the complexity of the language‟s syllable
inventory. 1 Computational methods involve the correlation of rhythm-types to acoustic
features such as the vocalic and consonantal intervals, one of which is the implementation
of Multivariate Discriminant Analysis (MDA). Perceptual methods involve contrasting the
rhythm of an unclassified language from prototype syllable-timed and stress-timed
sentences. In order to isolate rhythm from speech, a data-stripping technique called flat
sasasa resynthesis was implemented wherein the consonants are replaced with /s/ and
vowels with /a/, producing a resynthesized alternating “sasasa” sounds at a constant pitch
(F0). The rhythm discrimination and classification were closely examined for consistency
between the data modeling and listening test results. The computational experiment was
able to show that an MDA classifier trained to distinguish English and Japanese sentences
tend to label Filipino sentences as Japanese 67% of the time, vis-à-vis the perceptual
experiment showing that the listeners perceive Filipino to be more similar with Japanese,
this study shows computational and perceptual validation that Filipino is syllable-timed,
just like Japanese.
Patents by Timothy Israel Santos
Bottleneck Bandwidth Estimation Algorithm using Digital Signal Processing Techniques
Provided are a method of performing an event action by using a touch screen device based on opera... more Provided are a method of performing an event action by using a touch screen device based on operation gestures that are simultaneously input by using multiple operation tools and the touch screen device. The method includes: identifying a first operation tool based on contact by the first operation tool, the contact being sensed on the touch screen device; setting an operation area on the touch screen device based on an area designated by the contact by the first operation tool; identifying a second operation tool based on access by the second operation tool, the access being sensed on the touch screen device; sensing an operation gesture of the second operation tool within the operation area by using the second operation tool.
Uploads
Papers by Timothy Israel Santos
Education (MTBMLE) will require definitive rules for
orthography and grammar. While there are such rules for some
Philippine languages, there is a need to determine the agreement
and points of departure between the rules and the usage to avoid
confusion. This study makes an objective analysis of the levels of
agreement, in terms of grammar and orthographic rules, between
reference books and actual usage as evidenced from web-mined
text corpora for three major Philippine languages, namely
Filipino, Cebuano-Visayan and Ilokano. A list of language rules
on grammar and orthography were selected from standard
reference books for each of the aforementioned languages.
Alternative forms of usage for each selected language rule were
identified, and frequency counts were made, to be used as bases
for a comparative analysis between the rules prescribed by
standard reference books and actual language usage. The
techniques used in this study are important in language education,
serving to identify areas of confusion in language use in aspects of
grammar and orthography.
Filipino speech rhythm. Speech rhythm may be described as a language‟s distinguishing
durational sound pattern, resulting from the complexity of the language‟s syllable
inventory. 1 Computational methods involve the correlation of rhythm-types to acoustic
features such as the vocalic and consonantal intervals, one of which is the implementation
of Multivariate Discriminant Analysis (MDA). Perceptual methods involve contrasting the
rhythm of an unclassified language from prototype syllable-timed and stress-timed
sentences. In order to isolate rhythm from speech, a data-stripping technique called flat
sasasa resynthesis was implemented wherein the consonants are replaced with /s/ and
vowels with /a/, producing a resynthesized alternating “sasasa” sounds at a constant pitch
(F0). The rhythm discrimination and classification were closely examined for consistency
between the data modeling and listening test results. The computational experiment was
able to show that an MDA classifier trained to distinguish English and Japanese sentences
tend to label Filipino sentences as Japanese 67% of the time, vis-à-vis the perceptual
experiment showing that the listeners perceive Filipino to be more similar with Japanese,
this study shows computational and perceptual validation that Filipino is syllable-timed,
just like Japanese.
Patents by Timothy Israel Santos
Education (MTBMLE) will require definitive rules for
orthography and grammar. While there are such rules for some
Philippine languages, there is a need to determine the agreement
and points of departure between the rules and the usage to avoid
confusion. This study makes an objective analysis of the levels of
agreement, in terms of grammar and orthographic rules, between
reference books and actual usage as evidenced from web-mined
text corpora for three major Philippine languages, namely
Filipino, Cebuano-Visayan and Ilokano. A list of language rules
on grammar and orthography were selected from standard
reference books for each of the aforementioned languages.
Alternative forms of usage for each selected language rule were
identified, and frequency counts were made, to be used as bases
for a comparative analysis between the rules prescribed by
standard reference books and actual language usage. The
techniques used in this study are important in language education,
serving to identify areas of confusion in language use in aspects of
grammar and orthography.
Filipino speech rhythm. Speech rhythm may be described as a language‟s distinguishing
durational sound pattern, resulting from the complexity of the language‟s syllable
inventory. 1 Computational methods involve the correlation of rhythm-types to acoustic
features such as the vocalic and consonantal intervals, one of which is the implementation
of Multivariate Discriminant Analysis (MDA). Perceptual methods involve contrasting the
rhythm of an unclassified language from prototype syllable-timed and stress-timed
sentences. In order to isolate rhythm from speech, a data-stripping technique called flat
sasasa resynthesis was implemented wherein the consonants are replaced with /s/ and
vowels with /a/, producing a resynthesized alternating “sasasa” sounds at a constant pitch
(F0). The rhythm discrimination and classification were closely examined for consistency
between the data modeling and listening test results. The computational experiment was
able to show that an MDA classifier trained to distinguish English and Japanese sentences
tend to label Filipino sentences as Japanese 67% of the time, vis-à-vis the perceptual
experiment showing that the listeners perceive Filipino to be more similar with Japanese,
this study shows computational and perceptual validation that Filipino is syllable-timed,
just like Japanese.