Intelligent Camera Tracking using SRP-based sound Source localization in frequency domain
Independent thesis Advanced level (degree of Master (Two Years))Student thesis
The Steered Response Power Phase Transform (SRP-PHAT) is one of the most robust methods among sound source localization operating in noisy and reverberant environments. Direction of Arrival (DOA) Estimation has important applications in human computer interfaces such as video conferencing, speech enhancement and speech recognition. In this thesis work, SRP-PHAT method has been implemented for 16 element microphone array arranged into 4 rows and 4 columns in the presence of noise and reverberation. Computation of TDOA for each pair of microphones in a row setup or a column setup, generalized cross correlation estimates are calculated and thereby computing the source position and then by averaging the row wise obtained TDOA values and column wise obtained TDOA values, best accurate source position can be determined. Weighted Overlap and Add (WOLA) filter bank is used in SRP-PHAT method to find the TDOA in frequency domain. Original TDOA's and estimated TDOA's obtained from SRP-PHAT are compared to analyse the performance of the SRP-PHAT method. Mean estimation error and Standard deviation are calculated to find the accuracy of the estimated values of TDOA.
Place, publisher, year, edition, pages
2012. , 55 p.
SRP-Phat, acoustics, camera tracking, GCC Phat
Signal Processing Electrical Engineering, Electronic Engineering, Information Engineering
IdentifiersURN: urn:nbn:se:bth-4046Local ID: oai:bth.se:arkivexA2081AA51BDEA180C1257A9800222099OAI: oai:DiVA.org:bth-4046DiVA: diva2:831365
Grbic, Dr. NedelkoClaesson, Dr. Benny Sällberg & Dr. Ingvar