Sah Hj Salam

Sah Hj Salam

Auto-generated profile, based on publicly available information

Sah Hj Salam

Universiti Teknologi Malaysia

Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 March 2014 in IETE Technical Review
Verified
Journal Article
Closed access
JLJensen Wong Jing Lung
Jensen Wong Jing Lung
SSSah Hj. Salam
Sah Hj. Salam
ARAmjad Rehman
Amjad Rehman
MRMohd Shafry Mohd Rahim
Mohd Shafry Mohd Rahim
TSTanzila Saba
Tanzila Saba
The overall success of automatic speech recognition (ASR) depends on efficient phoneme recognition performance and quality of speech signal received in ASR. However, dissimilar inputs of speakers affect the overall recognition performance. One of the main problems that affect recognition performance is inter-speaker variability. Vocal tract length normalization (VTLN) is introduced to compensate inter-speaker variation on the speaker signal by applying speaker-specific warping of the frequency scale of a filter bank. Instead of measuring the performance on word level with speaker-specific warping, this research focuses on direct tackling at the phoneme level and applying VTLN on all speakers’ speech signals to analyse the best setting for the highest recognition performance. This research seeks to compare each phoneme recognition results from warping factor between 0.74 and 1.54 with 0.02 increments on nine different ranges of frequency warping boundary. The warp factor and frequency warping range that provides the highest phoneme recognition performance is applied on word recognition. The results show an improved performance in phoneme recognition by 0.7% and spoken word recognition by 0.5% using warp factor of 1.40 on frequency range of 300–5000 Hz in comparison to baseline results.

30 December 2022 in Electronics
Verified
Journal Article
Open access
PDF file
AAAmmar Mohammed Ali Alqadasi
Ammar Mohammed Ali Alqadasi
MSMohd Shahrizal Sunar
Mohd Shahrizal Sunar
Sherzod TuraevSherzod Turaev
Sherzod Turaev
RARawad Prof. Dr. Rawad Abdulghafor
Rawad Prof. Dr. Rawad Abdulghafor
SSSah Hj Salam
Sah Hj Salam
MAMohammed A. H. Ali
Mohammed A. H. Ali
Phoneme classification performance is a critical factor for the successful implementation of a speech recognition system. A mispronunciation of Arabic short vowels or long vowels can change the meaning of a complete sentence. However, correctly distinguishing phonemes with vowels in Quranic recitation (the Holy book of Muslims) is still a challenging problem even for state-of-the-art classification methods, where the duration of the phonemes is considered one of the important features in Quranic recitation, which is called Medd, which means that the phoneme lengthening is governed by strict rules. These features of recitation call for an additional classification of phonemes in Qur’anic recitation due to that the phonemes classification based on Arabic language characteristics is insufficient to recognize Tajweed rules, including the rules of Medd. This paper introduces a Rule-Based Phoneme Duration Algorithm to improve phoneme classification in Qur’anic recitation. The phonemes of the Qur’anic dataset contain 21 Ayats collected from 30 reciters and are carefully analyzed from a baseline HMM-based speech recognition model. Using the Hidden Markov Model with tied-state triphones, a set of phoneme classification models optimized based on duration is constructed and integrated into a Quranic phoneme classification method. The proposed algorithm achieved outstanding accuracy, ranging from 99.87% to 100% according to the Medd type. The obtained results of the proposed algorithm will contribute significantly to Qur’anic recitation recognition models.

1 April 2025
Verified
Book Chapter
Closed access
PDF file
SIShafina Ishigaki
Shafina Ishigaki
AIajune wanis ismail
ajune wanis ismail
NHNur Ameerah Abdul Halim
Nur Ameerah Abdul Halim
SSSah Hj Salam
Sah Hj Salam

17 February 2024 in Journal of Advanced Research in Computing and Applications
Verified
Journal Article
Open access
PDF file
SHSalam M S H
Salam M S H
JSJou T S
Jou T S
AFAhmad A F
Ahmad A F
Natural user interface (NUI) is an interface that enables users to interact with the digital world in the same way they interact with the physical world through sensory input such as touch, speech, and gesture. The combination of multiple modalities for NUI has become the trend in user interface recently. There is significant progress in advancing speech and hand recognition technology, which makes both become effective input modalities in HCI. However, there are limitations exist that degrade the performance which includes the complexity of vocabulary and unnatural hand gestures to instruct the machine. Therefore, this project aims to develop an application with natural gesture and speech input for 3D object manipulation. Three phases have been carried out, first is data collection and analysis, second is application structure design, and the third is the implementation of speech and gesture in 3D object manipulation. This application is developed by using Leap Motion Controller for hand gesture tracking, and Microsoft Azure Speech Cognitive Service and Microsoft Azure Language Understanding Intelligence Service for natural language speech recognition. The evaluation has been performed based on the accuracy of the command recognition and usability and user acceptance. The results show that the approaches developed in this project able to make good recognition of the speech command and gesture interaction while user experience testing shows high level of satisfaction in the application.

25 November 2024 in International Journal of Innovative Computing
Verified
Journal Article
Open access
PDF file
MYMohammed Yasin
Mohammed Yasin
SSSah Hj Salam
Sah Hj Salam
TATarmizi Adam
Tarmizi Adam
Active learning can help reduce the labeling burden in training deep object detection models by strategically selecting the most informative data points for labeling from unlabeled data. Different approaches to this task have been proposed in recent years, with differences often found in their approaches to the specific components. Consequently, this study breaks down active learning methods into their individual components and examines the diverse strategies associated with each intending to shed light on areas that have not yet been explored in depth. Doing so, the study offers valuable insights into potential research directions in this area.

22 June 2024
Verified
Journal Article
Open access
PDF file
AHAkm Hossain
Akm Hossain
SBSah Bin
Sah Bin
HSHj Salam
Hj Salam
MAMuhammad Alam
Muhammad Alam
AYAhmad Fadhil Yusof
Ahmad Fadhil Yusof
ABAkm Bellal
Akm Bellal
All body organs, especially the most important ones, are controlled by the brain, as well as any dysfunction of these cells would immediately jeopardize the life of a person by causing other organs to collapse. The brain is regarded as the most important organ in a person's body as a result. A tumor is a disorder of the brain's cells that manifests as inflamed brain tissue. The likelihood of quickly curing sickness will increase with the early identification of such tumorous cells. The detection of brain tumors is now done using magnetic resonance imaging (MRI). Several approaches have been created for MRI segmentation and tumor detection in the newly emerging study subject of image processing and segmentation of MRI images. An efficient threshold and region-based segmentation technique termed fuzzy c-means adaptive convolutional long/short-term memory (FCM-CLSTM) is proposed in this study, along with various morphological operations. To improve the MRI image quality, the following methods are first applied: normalized median filter (NMF), Histogram Equalization (HE), Contrast Limited Adaptive Histogram Equalization (CLAHE), and Brightness Preserving Dynamic Fuzzy Histogram Equalization (BPDFHE). The pixels are then divided into distinct classes using our suggested FCM-CLSTM approach's threshold and region-based segmentation, classification, and morphological operators, which are then used to find the tumor portion of the image with the highest intensity. The analysis is done on accuracy (88.24%), precision (87.66%), recall (87.66%), F1-score (87.66%), RMSE (0.3430), AMBE (0.0) and MSE (0.1176) with the existing methodologies. The findings of the proposed method achieved significant scores compared to the existing method in brain MRI segmentation.