Dithering techniques in automatic recognition of speech corrupted by MP3 compression: Analysis, solutions and experiments

2017 ◽  
Vol 86 ◽  
pp. 75-84 ◽  
Author(s):  
Michal Borsky ◽  
Petr Mizera ◽  
Petr Pollak ◽  
Jan Nouza
Keyword(s):  
Automatic Recognition ◽  
Mp3 Compression

Compressive Sensing Audio Watermarking dengan Metode LWT dan QIM

2018 ◽  
Vol 6 (3) ◽  
pp. 405
Author(s):  
IRMA SAFITRI ◽  
NUR IBRAHIM ◽  
HERLAMBANG YOGASWARA
Keyword(s):  
Wavelet Transform ◽  
Compressive Sensing ◽  
Audio Watermarking ◽  
Quantization Index Modulation ◽  
Computationally Efficient ◽  
Black And White ◽  
Additional Information ◽  
Audio Files ◽  
Index Modulation ◽  
Mp3 Compression

ABSTRAKPenelitian ini mengembangkan teknik Compressive Sensing (CS) untuk audio watermarking dengan metode Lifting Wavelet Transform (LWT) dan Quantization Index Modulation (QIM). LWT adalah salah satu teknik mendekomposisi sinyal menjadi 2 sub-band, yaitu sub-band low dan high. QIM adalah suatu metode yang efisien secara komputasi atau perhitungan watermarking dengan menggunakan informasi tambahan. Audio watermarking dilakukan menggunakan file audio dengan format *.wav berdurasi 10 detik dan menggunakan 4 genre musik, yaitu pop, classic, rock, dan metal. Watermark yang disisipkan berupa citra hitam putih dengan format *.bmp yang masing-masing berukuran 32x32 dan 64x64 pixel. Pengujian dilakukan dengan mengukur nilai SNR, ODG, BER, dan PSNR. Audio yang telah disisipkan watermark, diuji ketahanannya dengan diberikan 7 macam serangan berupa LPF, BPF, HPF, MP3 compression, noise, dan echo. Penelitian ini memiliki hasil optimal dengan nilai SNR 85,32 dB, ODG -8,34x10-11, BER 0, dan PSNR ∞.Kata kunci: Audio watermarking, QIM, LWT, Compressive Sensing. ABSTRACTThis research developed Compressive Sensing (CS) technique for audio watermarking using Wavelet Transform (LWT) and Quantization Index Modulation (QIM) methods. LWT is one technique to decompose the signal into 2 sub-bands, namely sub-band low and high. QIM is a computationally efficient method or watermarking calculation using additional information. Audio watermarking was done using audio files with *.wav format duration of 10 seconds and used 4 genres of music, namely pop, classic, rock, and metal. Watermark was inserted in the form of black and white image with *.bmp format each measuring 32x32 and 64x64 pixels. The test was done by measuring the value of SNR, ODG, BER, and PSNR. Audio that had been inserted watermark was tested its durability with given 7 kinds of attacks such as LPF, BPF, HPF, MP3 Compression, Noise, and Echo. This research had optimal result with SNR value of 85.32 dB, ODG value of -8.34x10-11, BER value of 0, and PSNR value of ∞.Keywords: Audio watermarking, QIM, LWT, Compressive Sensing.

scholarly journals Automatic Recognition of Fetal Facial Standard Plane in Ultrasound Image via Fisher Vector

PLoS ONE ◽  
2015 ◽  
Vol 10 (5) ◽  
pp. e0121838 ◽  
Author(s):  
Baiying Lei ◽  
Ee-Leng Tan ◽  
Siping Chen ◽  
Liu Zhuo ◽  
Shengli Li ◽  
...  
Keyword(s):  
Ultrasound Image ◽  
Automatic Recognition ◽  
Fisher Vector

scholarly journals Smart Care Using a DNN-Based Approach for Activities of Daily Living (ADL) Recognition

2020 ◽  
Vol 11 (1) ◽  
pp. 10
Author(s):  
Muchun Su ◽  
Diana Wahyu Hayati ◽  
Shaowu Tseng ◽  
Jiehhaur Chen ◽  
Hsihsien Wei
Keyword(s):  
Neural Network ◽  
Health Care ◽  
Activities Of Daily Living ◽  
Video Processing ◽  
Daily Living ◽  
Deep Neural Network ◽  
Recognition System ◽  
Automatic Recognition ◽  
Data Filtering ◽  
Accuracy Rate

Health care for independently living elders is more important than ever. Automatic recognition of their Activities of Daily Living (ADL) is the first step to solving the health care issues faced by seniors in an efficient way. The paper describes a Deep Neural Network (DNN)-based recognition system aimed at facilitating smart care, which combines ADL recognition, image/video processing, movement calculation, and DNN. An algorithm is developed for processing skeletal data, filtering noise, and pattern recognition for identification of the 10 most common ADL including standing, bending, squatting, sitting, eating, hand holding, hand raising, sitting plus drinking, standing plus drinking, and falling. The evaluation results show that this DNN-based system is suitable method for dealing with ADL recognition with an accuracy rate of over 95%. The findings support the feasibility of this system that is efficient enough for both practical and academic applications.

scholarly journals CAD integrated automatic recognition of weld paths

2021 ◽  
Author(s):  
Tuan Anh Tran ◽  
Andrei Lobov ◽  
Tord Hansen Kaasa ◽  
Morten Bjelland ◽  
Ole Terje Midling
Keyword(s):  
Life Cycle ◽  
Spatial Analysis ◽  
Product Life Cycle ◽  
Real Life ◽  
Automatic Recognition ◽  
Integrated Method ◽  
Weld Joints ◽  
Related Information ◽  
Almost All ◽  
Sequential Assembly

AbstractIn this paper, a CAD integrated method is proposed for automatic recognition of potential weld locations in large assembly structures predominantly comprised of weld joints. The intention is to reduce the total man-hours spent on manually locating, assigning, and maintaining weld-related information throughout the product life cycle. The method utilizes spatial analysis of extracted stereolithographic data in combination with available CAD functions to determine whether the accessibility surrounding a given intersection edge is sufficient for welding. To demonstrate the method, a system is developed in Siemens NX using their NXOpen Python API. The paper presents the application of the method to real-life use cases in varying complexity in cooperation with industrial partners. The system is able to correctly recognize almost all weld lines for the parts considered within a few minutes. Some exceptions are known for particular intersection lines located deep within notched joints and geometries weldable through sequential assembly, which are left as a subject to further works.

Sign in / Sign up

Export Citation Format

Share Document