Non-contact Quantification of Jugular Venous Pulse Waveforms from Skin Displacements

Figures

• 

  Figure 1. A normal ECG trace (on the top) and Jugular venous (JV) pressure waveforms (on the bottom). In a normal JV pressure trace the R peak of the ECG signal happens between the a peak and the c peak of the JV waveform. The waveforms are adapted from3,4,8.

• 

  Figure 2. The camera rig used to record the videos from the JV pulsation of the subjects. The two movable arms helps to adjust the camera position relative to the neck. A blue LED was used to illuminate the location of the recordings.

• 

  Figure 3. The typical position of the patient is 45° when evaluating jugular venous (JV) pressure in clinical examinations. This angle simplifies the process of estimating the blood pressure waveforms in the traditional method of using visual observations.

• 

  Figure 4. The experimental set-up used to record videos from the JV pulsation area. A laser displacement sensor was used to validate the displacement measurements of our method. The camera and the laser displacement measurement sensor were placed at approximately 45° with respect to the skin surface, and perpendicular to each other to avoid perspective distortion effects in the camera view.

• 

  Figure 5. Example frame showing three subimages and their associated control points (circles). The neck covers approximately two-thirds of the image. The region of interested is cropped and zoomed on the right image for visualisation purposes. The bright spot visible in the images arises from the laser displacement sensor.

• 

  Figure 6. A system diagram showing the four main steps of our method for measuring the JV displacement waveforms.

• 

  Figure 7. A displacement waveform and its fast Fourier transform (FFT). The maximum power for this waveform was at a frequency of 1.32 Hz. The top 5 % frequencies were used to filter the waveforms, which are more likely to be due to the JV pulsation.

• 

  Figure 8. Power amplitude (normalised) at tone frequency (between 0.7 Hz and 2 Hz) of control points close to the jugular vein. The red and green dots that have larger amplitudes compared to the rest of the measurements shown the region of JV pulsation in this recording.

• 

  Figure 9. Jugular venous pulse displacement waveforms over two cardiac cycles for each of 19 healthy subjects. The three important peaks of the JV waveform (a, c, and v) were observed in all the subjects. The second cardiac cycle of Subject 13 was the only waveform that did not have one of the peaks (the a peak is missing).

• 

  Figure 10. Normalised jugular vein displacement waveforms measured using our method (black lines) and using a laser displacement sensor (orange lines) derived from two recordings in each of three subjects. The correlation scores between the associated pairs of recordings in panels A to F were 0.98, 0.97, 0.96, 0.99, 0.96, and 0.95, respectively.

• 

  Figure 11. Normalised jugular vein displacement waveforms measured using our method, and synchronised with recorded ECG traces for two recordings in each of three subjects. The R peak of the ECG signal has happened close to the a peak of the JV waveform.

• 

  Figure 12. Measured JV waveforms from video recordings for five cardiac cycles of subject 11. The video recordings were synchronised with ECG signals. To perform a comparison between the recordings, two time intervals were defined. The first interval was the time between the v descent of the JV waveforms and the R peak of the ECG signal (v-R), and the second one was the duration between the R peak of the ECG signal and the c peak of the JV waveforms. All the durations were normalized for one cycle duration (v to v).