Estimation of vital signs from facial videos via video magnification and deep learning.

The continuous monitoring of vital signs is one of the hottest topics in healthcare. Recent technological advances in sensors, signal processing, and image processing spawned the development of no-contact techniques such as remote photoplethysmography (rPPG). To solve the common problems of rPPG including weak extracted signals, body movements, and generalization with limited data resources, we proposed a dual-path estimation method based on video magnification and deep learning. First, image processes are applied to detect, track, and magnificate facial ROIs automatically. Then, the steady part of the wave of each processed ROI is used for the extraction of features including heart rate, PTT, and features of pulse wave waveform. The blood pressures are estimated from the features via a small CNN. Results comply with the current standard and promise potential clinical applications in the future.

Three-Dimensional Segmentation and Quantitative Measurement of the Aqueous Outflow System of Intact Mouse Eyes Based on Spectral Two-Photon Microscopy Techniques.

PURPOSE: To visualize and quantify the three-dimensional (3D) spatial relationships of the structures of the aqueous outflow system (AOS) within intact enucleated mouse eyes using spectral two-photon microscopy (TPM) techniques. METHODS: Spectral TPM, including two-photon autofluorescence (TPAF) and second-harmonic generation (SHG), were used to image the small structures of the AOS within the limbal region of freshly enucleated mouse eyes. Long infrared excitation wavelengths (930 nm) were used to reduce optical scattering and autofluorescent background. Image stacks were collected for 3D image rendering and structural segmentation. For anatomical reference, vascular perfusion with fluorescent-conjugated dextran (150 KDa) and trans-corneal perfusion with 0.1 µm fluorescent polystyrene beads were separately performed to identify the episcleral veins (EV) and the trabecular meshwork (TM) and Schlemm's canal (SC), respectively. RESULTS: Three-dimensional rendering and segmentation of spectral two-photon images revealed detailed structures of the AOS, including SC, collector channels (CC), and aqueous veins (AV). The collagen of the TM was detected proximal to SC. The long and short axes of the SC were 82.2 ± 22.2 µm and 6.7 ± 1.4 µm. The diameters of the CC averaged 25.6 ± 7.9 µm where they originated from the SC (ostia), enlarged to 34.1 ± 13.1 µm at the midpoint, and narrowed to 18.3 ± 4.8 µm at the junction of the AV. The diameter of the AV averaged 12.5 ± 3.4 µm. CONCLUSIONS: Spectral TPM can be used to reconstruct and measure the spatial relationships of both large and small AOS structures, which will allow for better understanding of distal aqueous outflow dynamics.

Gold nanoaggregates for probing single-living cell based on surface-enhanced Raman spectroscopy.

Gold nanoparticles are delivered into living cells by transient electroporation method to obtain intracellular surface-enhanced Raman spectroscopy (SERS). The subcellular localization of gold nanoparticles is characterized by transmission electron microscopy, and the forming large gold nanoaggregates are mostly found in the cytoplasm. The SERS detection of cells indicates that this kind of gold nanostructures induces a high signal enhancement of cellular chemical compositions, in addition to less cellular toxicity than that of silver nanoparticles. These results demonstrate that rapid incorporation of gold nanoparticles by electroporation into cells has great potential applications in the studies of cell biology and biomedicine.

Label-free detection of serum proteins using surface-enhanced Raman spectroscopy for colorectal cancer screening.

Surface-enhanced Raman scattering (SERS) spectra of serum proteins purified from human serum samples were employed to detect colorectal cancer. Acetic acid as a new aggregating agent was introduced to increase the magnitude of the SERS enhancement. High-quality SERS spectra of serum proteins were acquired from 103 cancer patients and 103 healthy volunteers. Tentative assignments of SERS bands reflect that some specific biomolecular contents and protein secondary structures change with colorectal cancer progression. Principal component analysis combined with linear discriminant analysis was used to assess the capability of this approach for identifying colorectal cancer, yielding diagnostic accuracies of 100% (sensitivity: 100%; specificity: 100%) based on albumin SERS spectroscopy and 99.5% (sensitivity: 100%; specificity: 99%) based on globulin SERS spectroscopy, respectively. A partial least squares (PLS) approach was introduced to develop diagnostic models. An albumin PLS model successfully predicted the unidentified subjects with a diagnostic accuracy of 93.5% (sensitivity: 95.6%; specificity: 91.3%) and the globulin PLS model gave a diagnostic accuracy of 93.5% (sensitivity: 91.3%; specificity: 95.6%). These results suggest that serum protein SERS spectroscopy can be a sensitive and clinically powerful means for colorectal cancer detection.

Label-free detection of blood plasma using silver nanoparticle based surface-enhanced Raman spectroscopy for esophageal cancer screening.

A surface-enhanced Raman spectroscopy (SERS) based on silver nanoparticle technology was applied to analyze and classify human blood plasma with the aim to develop a simple and label-free blood test for esophageal cancer detection. High quality SERS spectra in the range of 400-1800 cm(-1) can be acquired from 36 esophageal cancer patients and 50 healthy volunteers' blood plasma samples. Tentative assignments of the SERS bands indicated specific biomolecular changes associated with cancer transformation, including an increase in the relative amounts of nucleic acid and phenylalanine, a decrease in the percentage of saccharide and proteins contents in the cancer blood plasma compared to that of healthy subjects. Furthermore, both SVM and PCA-LDA diagnostic algorithm were employed to analyze and classify the obtained blood plasma SERS spectra between normal and cancer plasma with a high diagnostic accuracy (around 90%). This exploratory work demonstrates that the label-free plasma SERS analysis technique in conjunction with SVM and PCA-LDA diagnostic algorithms has great potential for improving esophageal cancer detection and screening.

Label-free imaging characteristics of colonic mucinous adenocarcinoma using multiphoton microscopy.

Colorectal carcinoma (CRC) has high mortality and increased incidence rates. An early detection of CRC is very important. Multiphoton microscopy (MPM) with high resolution and high sensitivity is used to effectively distinguish the microstructure changes of normal and mucinous adenocarcinoma slices of ex vivo human colonic tissues. In mucinous adenocarcinoma mucosa, the glands are distorted and elongated, the gland cavity is indistinct, and the mesh collagen fibers are diminished. In the submucosa, the collagens are seriously disordered, elongated, pushed aside, and sparsely visible, the content of elastic fibers is also broken and almost disappearing. Many cancer cells, some in cavity-like shape full of mucus surrounded by some collagen fibers, occupied the submucosa, which are comparable to hematoxylin-eosin (HE) stained images. Second harmonic generation and two-photon excitation fluorescence (SHG/TPEF) intensity ratio can be used further to quantitatively evaluate normality and abnormality. The fast Fourier transform (FFT) images show that the normal collagen fibrils are dense and in random order, and the cancerous collagen is certainly organized. The exploratory results show that it has potential for the development of multiphoton mini-endoscopy in real-time early diagnosis of CRC.

Establishing diagnostic features for identifying the mucosa and submucosa of normal and cancerous gastric tissues by multiphoton microscopy.

BACKGROUND: Establishing diagnostic features is essential and significant for developing multiphoton endoscopy to make an early diagnosis of gastric cancer at the cellular level. Until now, these diagnostic features have not been clearly described and understood. DESIGN: Study of diagnostic features based on multiphoton microscopy (MPM). OBJECTIVE: Establishing diagnostic features to identify the mucosa and submucosa of human normal and cancerous gastric tissues by investigating their multiphoton microscopic images. SETTING: Fujian Normal University and Fujian Provincial Tumor Hospital. PATIENTS: Ten pairs of normal and cancerous specimens were obtained from 10 patients (ages 51-68 years) undergoing radical gastrectomy. INTERVENTIONS: MPM was performed on specimens. MAIN OUTCOME MEASUREMENTS: Establishment of diagnostic features. RESULTS: MPM has the ability to exhibit not only the mucosal and submucosal microstructures of normal and cancerous gastric tissues but also the distribution and content of abnormal cells in these 2 layers. More importantly, it can provide the diagnostic features to qualitatively and quantitatively differentiate between normal and cancerous gastric tissues. LIMITATIONS: The selection bias and preparation of specimen. CONCLUSIONS: These findings provide the groundwork for further establishing diagnostic criteria.

A pilot study of using multiphoton microscopy to diagnose gastric cancer.

BACKGROUND: Using a combination of autofluorescence from cells and second-harmonic generation (SHG) signal from collagen, multiphoton microscopy (MPM) imaging can provide detailed real-time information on tissue architecture and cellular morphology in live tissue without administration of exogenous contrast agents. The purpose of this study is to evaluate the feasibility of using MPM to histologically diagnose gastric cancer by using fresh, unfixed, unstained gastric specimens, compared with gold-standard hematoxylin-eosin (H-E)-stained histopathology. METHODS: A pilot study was performed between June 2009 and December 2009. Ten cases with gastric carcinoma confirmed by preoperative endoscopic biopsy underwent radical gastrectomy. The fresh specimen was opened, and a piece of cancer tissue and a piece of normal tissue each with a size of 1-1.5 cm across and 0.2 cm in thickness were taken and snap-frozen. A 5-µm slide was sectioned for MPM examination, and the remainder of the tissue went through routine histopathological procedure. MPM images and H-E-stained images were compared by the same attending pathologist. RESULTS: MPM images were acquired by two channels: broadband autofluorescence from cells, and SHG from tissue collagen. Peak multiphoton autofluorescence intensity was detected in mucosa excited at 800 nm. Cancer cells, characterized by irregular size and shape, enlarged nuclei, and increased nuclear-to-cytoplasmic ratio, were identified by MPM images, which were confirmed by H-E-stained images. Regular architectures of gastric pits and gastric glands in the normal tissue of the same specimens were clearly revealed by MPM images, which were comparable to H-E-stained images. CONCLUSIONS: It is feasible to use MPM to diagnose gastric cancer by "optical biopsy." With miniaturization and integration of endoscopy, MPM has the potential to provide real-time histological diagnosis without invasive biopsy for gastric cancer in the future.

Multiphoton microscopic imaging of normal human rectum tissue.

In this paper, multiphoton microscopy (MPM), based on two-photon excited fluorescence and second harmonic generation signals, was used to image microstructures of human rectal mucosa and submucosa. The morphology and distribution of the main components in mucosa layer, goblet cells, intestinal glands, and a little collagen fibers have been clearly monitored, and the content and distribution of collagen, elastic fibers, and blood vessels in submucosa layer have also been distinctly obtained. The variation of these components is very relevant to the pathology in gastrointestinal system, especially early rectal cancer. Our results indicate that the MPM technique has the potential application in vivo in the clinical diagnosis and monitoring of early rectal cancer.