Impact of random and scattered coincidences from outside of field of view on positron emission tomography/computed tomography imaging with different reconstruction protocols

Image quality in positron emission tomography(PET)is affected by random and scattered coincidences and reconstruction protocols.In this study,we investigated the effects of scattered and random coincidences from outside the field of view(FOV)on PET image quality for different reconstruction protocols.Imaging was performed on the Discovery 690 PET/CT scanner,using experimental configurations including the NEMA phantom(a body phantom,with six spheres of different sizes)with a signal background ratio of 4:1.The NEMA phantom(phantom I)was scanned separately in a one-bed position.To simulate the effect of random and scatter coincidences from outside the FOV,six cylindrical phantoms with various diameters were added to the NEMA phantom(phantom II).The 18 emission datasets with mean intervals of 15 min were acquired(3 min/scan).The emission data were reconstructed using different techniques.The image quality parameters were evaluated by both phantoms.Variations in the signal-to-noise ratio(SNR)in a 28-mm(10-mm)sphere of phantom II were 37.9%(86.5%)for ordered-subset expectation maximization(OSEM-only),36.8%(81.5%)for point spread function(PSF),32.7%(80.7%)for time of flight(TOF),and 31.5%(77.8%)for OSEM+PSF+TOF,respectively,indicating that OSEM+PSF+TOF reconstruction had the lowest noise levels and lowest coefficient of variation(COV)values.Random and scatter coincidences from outside the FOV induced lower SNR,lower contrast,and higher COV values,indicating image deterioration and significantly impacting smaller sphere sizes.Amongst reconstruction protocols,OSEM+PSF+TOF and OSEM+PSF showed higher contrast values for sphere sizes of 22,28,and 37 mm and higher contrast recovery coefficient values for smaller sphere sizes of 10 and 13 mm.

Word perception in noise at different channels in simulated cochlear implant listeners

AIM:To find out effect of different signal-to-noise ratios(SNRs) on word perception at different number of channels.METHODS: Thirty participants with normal hearing in the age range of 18-25 years(mean age 23.6 years) were involved in the study. For word perception test, there were 28 key-words embedded in sentences comprises of four lists processed for different channels(4, 8 and 32 channel) using AngelS im program at-5, 0 and +5 SNRs. The recorded stimuli were routed through audiometer connected with computer with CD player and presented in free field condition with speakers kept at 0° azimuth in a sound treated room.RESULTS: Repeated measure ANOVA showed significant main effect across different SNRs at 4 channel, 8 channel and at 32 channel. Further, Bonferroni multiple pairwise comparisons shows significant differences between all the possible combinations(4, 8 and 32 channel) at +5 dB SNR, 0 dB SNR and-5 dB SNR.CONCLUSION: Present study highlights the importance of more number of channels and higher signal to noise ratio for better perception of words in noise in simulated cochlear implantees.

Ghost edge detection based on HED network

In this paper,we present an edge detection scheme based on ghost imaging(GI)with a holistically-nested neural network.The so-called holistically-nested edge detection(HED)network is adopted to combine the fully convolutional neural network(CNN)with deep supervision to learn image edges efectively.Simulated data are used to train the HED network,and the unknown object’s edge information is reconstructed from the experimental data.The experiment results show that,when the compression ratio(CR)is 12.5%,this scheme can obtain a high-quality edge information with a sub-Nyquist sampling ratio and has a better performance than those using speckle-shifting GI(SSGI),compressed ghost edge imaging(CGEI)and subpixel-shifted GI(SPSGI).Indeed,the proposed scheme can have a good signal-to-noise ratio performance even if the sub-Nyquist sampling ratio is greater than 5.45%.Since the HED network is trained by numerical simulations before the experiment,this proposed method provides a promising way for achieving edge detection with small measurement times and low time cost.