Measuring heart rate traditionally requires special equipment and physical contact with the subject. Reliable non-contact and low-cost measurements are highly desirable for convenient and comfortable physiological self-assessment. Previous work has shown that consumer-grade cameras can provide useful signals for remote heart rate measurements. In this paper a simple and robust method of measuring the heart rate using low-cost webcam is proposed. Blood volume pulse is extracted by proper Region of Interest (ROI) and color channel selection from image sequences of human faces without complex computation. Heart rate is subsequently quantified by spectrum analysis. The method is successfully applied under natural lighting conditions. Results of experiments show that it takes less time, is much simpler, and has similar accuracy to the previously published and widely used method of Independent Component Analysis (ICA). Benefitting from non-contact, convenience, and low-costs, it provides great promise for popularization of home healthcare and can further be applied to biomedical research.
Non-contact and remote measurements of vital physical signals are important for reliable and comfortable physiological
self-assessment. In this paper, we provide a new video-based methodology for remote and fast measurements of vital
physical signals such as cardiac pulse and breathing rate. A webcam is used to track color video of a human face or wrist,
and a Photoplethysmography (PPG) technique is applied to perform the measurements of the vital signals. A novel
sequential blind signal extraction methodology is applied to the color video under normal lighting conditions, based on
correlation analysis between the green trace and the source signals. The approach is successfully applied in the
measurement of vital signals under the condition of different illuminating in which the target signal can also be found out
accurately. To assess the advantages, the measuring time of a large number of cases is recorded correctly. The
experimental results show that it only takes less than 30 seconds to measure the vital physical signals using presented
technique. The study indicates the proposed approach is feasible for PPG technique, which provides a way to study the
relationship of the signal for different ROI in future research.
A method to accurately measure landmarks of Hanjiangfish fossil based on computer vision is presented. Hanjiangfish is
a deracinated vertebrate ever lived on the earth four million years ago. The geometric morphometry of the ichthyolite
based on the landmark measurement is fundamental in the vertebrate palaeontology. The landmark measurement is based
on digital model of the fossil. The Hanjiangfish fossil is measured fast and accurately by using the structured-light
scanning method. The digital model of the fossil then is made from the cloudy data through the acquisition, align, merge,
and edit processes. The local least-square method is used to fit two lines near the measured landmark. The landmark is
the point of intersection of the two fitting lines, which is unique and accurate. The distance between all pairs of
landmarks and interior angles from a triangulation of the landmarks can be measured fast and accurately by the digital
model. Other important geometric parameters of the fossil, such as curvature and surface area can also be measured by
the digital model. The Hanjiangfish fossil is also measured accurately by a universal tool microscope, which is used as a
measuring standard with higher accuracy to validate the proposed method. The experiment shows that the measured
errors of the distance and angle are less than 0.1 mm and 1' respectively, which are good enough for geometric
morphometry of palaeontology.
The landmark measurement of human skull is fundamental to geometric morphometry of palaeoanthropology. The
landmarks are geometry points which can describe anatomically the homology of species group. They play an important
role in palaeoanthropology. A structured-light based method is used to measure and make the 3D digital model of skull.
The distances between all pairs of landmarks and interior angles from triangulations of the landmarks can be measured
fast and accurately by the digital model. Other important geometric parameters of the skull, such as curvature, surface
area, volume can also be measured. In order to validate and certificate the proposed method, 9 standard balls, which are
embed at the landmarks, are measured by using Coordinate Measuring Arm (CMA). The experiment shows that the
measuring errors of the distances and angles are less than 0.08 mm and 5' respectively.
Osteometry is fundamental to study the human skeleton. It has been widely used in palaeoanthropology, bionics, and
criminal investigation for more than 200 years. The traditional osteometry is a simple 1-dimensional measurement that
can only get 1D size of the bones in manual step-by-step way, even though there are more than 400 parameters to be
measured. For today's research and application it is significant and necessary to develop an advanced 3-dimensional
osteometry technique. In this paper a new 3D osteometry is presented, which focuses on measurement of the femur, the
largest tubular bone in human body. 3D measurement based on the structured light scanning is developed to create fast
and precise measurement of the entire body of the femur. The cloud data and geometry model of the sample femur is
established in mathematic, accurate and fast way. More than 30 parameters are measured and compared with each other.
The experiment shows that the proposed method can meet traditional osteometry and obtain all 1D geometric parameters
of the bone at the same time by the mathematics model, such as trochanter-lateral condyle length, superior breadth of
shaft, and collo-diaphyseal angle, etc. In the best way, many important geometric parameters that are very difficult to
measure by existing osteometry, such as volume, surface area, and curvature of the bone, can be obtained very easily.
The overall measuring error is less than 0.1mm.
Osteometry is fundamental for anthropometry. It provides the key technology and value to the study of
palaeoanthropology, medicine, and criminal investigation. The traditional osteometry that has been widely accepted and
used since 18th century has no longer met the information demand for modern research and application. It is significant
and necessary to create an advanced 3-dimensional osteometry technique for anthropometry. This paper presents a new
quick and accurate method to measure human pelvis through mathematical modeling. The pelvis is a complex
combination of bones, which consists of three connected parts: hipbones, sacrum, and coccyx. There are over 40 items to
be measured for the 1-dimension characteristics. In this paper, a combined measuring technology is developed for pelvis
measurement. It uses machine vision systems and a portable measuring arm to obtain key geometry parameters of the
pelvis. The mathematics models of the pelvis spatial structure and its parts are created through the process of data
collecting, digging, assembling, and modeling. The experiment shows that the proposed technology can meet traditional
osteometry and obtain entire 1D geometric parameters of the pelvis, such as maximum breadth and height, diameter of
obstetric conjugata, inclination angle, and sakralneigungswinkel, etc. at the same time after modeling. Besides making
the measurements above, the proposed technology can measure the geometry characteristics of pelvis and its parts, such
as volume, surface area, curvature, and spatial structure, which are almost impossible for traditional technology. The
overall measuring error is less than 0.1mm.
A Laser Guiding Measuring Robot (LGMR) based on the new technology of Laser-Guiding, SMR-Tracking has been
developed. LGMR can be guided by measuring laser beam to do 3D laser tracking measurement automatically. LGMR
consists of a measuring robot and a laser tracker system (LTS). The measuring robot is employed to carry SMR to track the
measuring laser beam from LTS. LTS is used to measure 3D position of SMR and then complete the measurement. The
CAD model of a measured object can be used to control the measuring laser beam from LTS to point to the measured
position. The measuring robot then tarcks the guiding laser beam and drives SMR to the measured position. This paper
presents the working principle and system framework of LGMR. The details of the robot design, implement and
experiment are also provided. The experiments prove that the proposed LGMR can measure a complicated object
automatically by using the CAD model of the measured objects. The developed LGMR makes it possible for LTS to do 3D
tracking measurement automatically by using the CAD model of a measured object to guide the measuring robot.
Basic clinical diagnosis principles are based on the application of neural networks embedded in the expert system of cerebral potential signal. Analysis of ordinary expert system, development condition and main property of neural networks is provided, in the meantime argumentation of recognition, acquisition, fundamental structure design and intelligent diagnosis is also available in order to explain the effectiveness and advancement of our method.