Application Cases

Experiment Name: Application of Preamplifier for Weak Signals in Blood Vessel Identification Research Using Photoacoustic Technology

Research Direction: Biometric Identification Technology

Test Objective:
Utilize MATLAB for photoacoustic blood vessel identification: 1. Preprocess images from the photoacoustic blood vessel library, including normalization, binarization, smoothing, thinning, and burr trimming to obtain thinned images; 2. Extract feature values of the thinned vessel images using seven moment invariants to establish a dataset; 3. Use SVM to identify vessel images, confirming the feasibility of photoacoustic-based vessel image recognition.

Testing Equipment: ATA-5620 Preamplifier for Weak Signals, Solid-State Laser, Ultrasonic Transducer, Mechanical Displacement Platform, Oscilloscope, Filter, Weak Signal Amplifier, Computer, Data Acquisition Card

Experimental Procedure:
An OR-PAM system based on a solid-state laser was constructed using the testing equipment. An optical parametric oscillator (OPO) (OPOletteTM 532, Opotek LLC) was used as the light source: wavelength, pulse frequency, and pulse duration were 532 nm, 20 Hz, and 7 ns, respectively. The laser output power was controlled by a computer. After spatial attenuation and shaping, the OPO laser output beam was coupled into an optical fiber via a custom fiber coupler. The output assembly of the fiber was connected to an optical lens barrel. The optical lens barrel contained a fiber collimator with a focal length of 30 mm and a focusing lens with a focal length of 5 mm. The spot diameter was approximately 20 μm, suitable for imaging larger diameter main blood vessels, while smaller diameter side vessels were filtered out during subsequent image preprocessing. Subsequently, the optical lens was mounted on a three-dimensional mechanical displacement platform via a fixture for scanning. Signals were acquired by a non-focused ultrasonic receiver with a center frequency of 2.5 MHz. The acoustic signals were amplified by approximately 60 dB using the ATA-5620 preamplifier for weak signals. A multifunction data acquisition card was used to convert and acquire the analog signals, which were then digitized by a personal computer. The sample platform consisted of the displacement platform and the ultrasonic transducer.

Figure: (a) Schematic diagram of the photoacoustic blood vessel acquisition system structure; (b) Photoacoustic imaging system

During the experiment, 40 artificial blood vessels were first prepared as samples. Photoacoustic detection was performed on the vessel samples. Anti-counterfeiting was conducted based on the amplitude, velocity, and difference in arrival time of the signal maximum of the photoacoustic signal to distinguish counterfeit vessels. Subsequently, samples that met the photoacoustic characteristics (i.e., samples identified as genuine blood vessels) underwent photoacoustic imaging, and an image library was established. The photoacoustic vessel images, after image preprocessing, were then thinned and subjected to feature matching. Finally, the feature vector set was used as a dataset input into the SVM for classification and identification.

Experimental Results:

Figure: Photoacoustic Signals of Genuine and Fake Vessel Images

1. Amplitude
Comparing the peak-to-peak values of the photoacoustic signal used as a detection parameter, the figure shows the photoacoustic signals of fake vessels and genuine vasculature under the same energy. The amplitude of the fake vessels is significantly smaller than that of the genuine vessels.

2. Depth
Depth of the detected target: Depth = (Difference in maximum arrival time) × (Ultrasound velocity). During the experiment, the relative positions of the laser source, sample, and ultrasonic transducer remained constant.

Figure: Photoacoustic Signal of Vasculature After Using a Refractive Index Plate

After altering the final energy of the laser beam using a refractive index plate, the signal from the genuine vessels changed. The peak-to-peak values of the signals from fake and genuine vessels were 188 mV and 182.9 mV, respectively. The error between the peak-to-peak values of the fake and genuine vessel images was only (188-182.9)/182.9 ≈ 2.79%, which is negligible from a physical perspective. The maximum arrival time of the fake vessel signal was 7.042 µs.

3. Photoacoustic Vessel Identification

Figure: Genuine Vessel Images

Figures (a–f) show the genuine vessel image, and the vessel images after normalization (by size and grayscale), binarization, smoothing, thinning, and burr trimming, respectively. Features from each vessel image were extracted to form a dataset.

This dataset contained 160 vessel images from 10 rats, with 80 images used as the training set and the remaining images as the test set. The results of vessel identification are shown in Table 3. Using the SVM method, the recognition rate reached 97.5%. Therefore, the actual recognition rate of this system was 97.5%. Compared to results obtained by directly using SVM, it improved the recognition rate for distinguishing genuine and fake vessels by 2.63% ((97.5%-95%)/95%), significantly enhancing the recognition rate of the vessel identification system.