Table 1 Comparative Study of state-of-the art related works.

Huo et al.¹⁴

Used optical fiber to transmit fingertip data to the spectrometers

used Finger-end transmission method for spectral measurement

used two scientific research-grade ultra-high-sensitivity spectrometers

Dataset

Correlation coefficient

RMSE

Training set

0.93

0.57 × 10⁹

Prediction set

0.82

0.83 × 10⁹

Bourquard et al.²

Introduced a semi-automatic method to acquire video data with low-cost portable microscopy equipment

Performed Spatiotemporal analysis of capillary profiles

Classified severe neutropenia versus baseline neutropenia

The classification performance improved with the number of capillaries used per patient, achieving AUCs of 0.68, 0.84, 0.88, 0.95, and 1.00 for one to five capillaries, respectively

Bourquard et al.¹⁷

Used portable, commercially available capillaroscope to deduced to estimate the total sampled blood volume and identified all WBC events occurring in two capillaries of one subject and estimated their speed

The speed of WBC events was estimated from visual gap trajectories, falling within the known range for nailfold capillary blood flow. Based on the healthy WBC range and video duration, the predicted WBC counts were consistent with the observed median counts in both capillaries

Bagramyan et al.¹⁸

Performed non-invasive and label-free temporal recordings of circulating, rolling, and adherent leukocytes within the human buccal microvascular

Visualized fine morphological features of blood cells through phase contrast signal

Used an Oral tissue stabilization apparatus

Calculated average rolling velocity of 58 ± 28 µm/s in the healthy tissue, which reduced to 4 ± 6 µm/s in the inflamed tissue