The use of cameras and computational algorithms for noninvasive, low-cost and scalable measurement of physiological (e.g., cardiac and pulmonary) vital signs is very attractive. However, diverse data representing a range of environments, body motions, illumination conditions and physiological states is laborious, time consuming and expensive to obtain. Synthetic data have proven a valuable tool in several areas of machine learning, yet are not widely available for camera measurement of physiological states. Synthetic data offer perfect labels (e.g., without noise and with precise synchronization), labels that may not be possible to obtain otherwise (e.g., precise pixel level segmentation maps) and provide a high degree of control over variation and diversity in the dataset. We present SCAMPS, a dataset of synthetics containing 2,800 videos (1.68M frames) with aligned cardiac and respiratory signals and facial action intensities. The RGB frames are provided alongside segmentation maps and precise descriptive statistics about the underlying waveforms, including inter-beat interval, heart rate variability, and pulse arrival time. Finally, we present baseline results training on these synthetic data and testing on real-world datasets to illustrate generalizability.

The use of cameras and computational algorithms for noninvasive, low-cost and scalable measurement of physiological (e.g., cardiac and pulmonary) vital signs is very attractive. However, diverse data representing a range of environments, body motions, illumination conditions and physiological states is laborious, time consuming and expensive to obtain. Synthetic data have proven a valuable tool in several areas of machine learning, yet are not widely available for camera measurement of physiological states. Synthetic data offer "perfect" labels (e.g., without noise and with precise synchronization), labels that may not be possible to obtain otherwise (e.g., precise pixel level segmentation maps) and provide a high degree of control over variation and diversity in the dataset. We present SCAMPS, a dataset of synthetics containing 2,800 videos (1.68M frames) with aligned cardiac and respiratory signals and facial action intensities.

Selective clustering annotated using modes of projections (SCAMP) is a new clustering algorithm for data in $\mathbb{R}^p$. SCAMP is motivated from the point of view of non-parametric mixture modeling. Rather than maximizing a classification likelihood to determine cluster assignments, SCAMP casts clustering as a search and selection problem. One consequence of this problem formulation is that the number of clusters is $\textbf{not}$ a SCAMP tuning parameter. The search phase of SCAMP consists of finding sub-collections of the data matrix, called candidate clusters, that obey shape constraints along each coordinate projection. An extension of the dip test of Hartigan and Hartigan (1985) is developed to assist the search. Selection occurs by scoring each candidate cluster with a preference function that quantifies prior belief about the mixture composition. Clustering proceeds by selecting candidates to maximize their total preference score. SCAMP concludes by annotating each selected cluster with labels that describe how cluster-level statistics compare to certain dataset-level quantities. SCAMP can be run multiple times on a single data matrix. Comparison of annotations obtained across iterations provides a measure of clustering uncertainty. Simulation studies and applications to real data are considered. A C++ implementation with R interface is $\href{https://github.com/RGLab/scamp}{available\ online}$.

Researchers at Stanford University have developed a spider-like drone that can cling to walls and even perch on the ceiling. The versatile drone is equipped with'micro-spines' which create an opposing grip, allowing it to sit on rough, outdoor surfaces. These capabilities have potential applications in the monitoring of typically hard-to-reach areas, and its uses could range anywhere from the detection of damage on bridges to assisting in rescue missions. Researchers at Stanford University have developed a spider-like drone that can cling to walls and even perch on the ceiling. The versatile drone is equipped with'micro-spines' which create an opposing grip, allowing it to sit on rough, outdoor surfaces Micro-spines create two opposing grips, which pull inward to counterbalance the vehicle's movement and provide a grasping force.

Morgan Pope is a PhD student investigating robots that live at the boundary of airborne and surface locomotion at Stanford's Biomimetics and Dexterous Manipulation Lab. He wrote about SCAMP, a flying and perching robot, for Automaton earlier this year. These places have something in common: we have a need to understand what's going on where established infrastructure can't give us good data. Advances in computation, fabrication, and materials over the last half-century have resulted in small, cheap, and lightweight sensors that can provide us with these data; now the task is to find ways to deploy such sensors rapidly and effectively. One way to do this is with small, agile aerial vehicles like quadrotors.

As far as practical applications, SCAMP is designed for outdoor work in places like earthquake zones, where there's no usable flat surface on which a drone could land. It's not uncommon for walls to remain standing amid rubble in disaster areas, and an appropriately equipped SCAMP could perch on one to track seismic activity, for example, or to serve as a link in an emergency communications network. Its climbing capability would also allow it to get to an optimal location, rather than attempting to reposition itself through flying--which can be especially difficult in inclement weather.

While it looks nothing more than an unassuming quadcopter, Stanford's SCAMP (Stanford Climbing and Aerial Maneuvering Platform) has a lot more tricks up it sleeve--this drone can not only fly, it can also perch, and climb on walls. SCAMP basically takes everything the Biomimetics and Dexterous Manipulation Lab has learned from previous projects, such as the Stickybot (which mimics the gecko's wall climbing capability), to create this new drone. The team modified the climbing technology applied on the Stickybot so that SCAMP could climb faster. To achieve SCAMP's current maneuverability, they ensured it could take longer steps and added microspines to its feet--similar to what a praying mantis has. To achieve its ability to perch, the climbing mechanism for the machine was placed on top of the quadrotor, which allows it to press against surfaces for better stability.

Morgan Pope is a Ph.D student investigating robots that live at the boundary of airborne and surface locomotion at Stanford's Biomimetics and Dextrous Manipulation Lab. He's the lead author on a paper about SCAMP that is in review for IEEE Transactions on Robotics, and enjoys reading, Star Wars, and trying to keep up with his three small children. What goes up must come down--unless it can perch on something first. Quadrotors have limited endurance because of restrictions on battery capacity and the physics of small-scale flight, but perching can allow them to operate for hours or even days, gathering data or performing communication tasks while stationary. Perching can be tricky, because the odds of your drone landing in just the right place are low.

A team of engineers has built a robot which can fly, land and scuttle up walls, just like a bug. The researchers said the machines could be used in disaster areas where rubble or floodwaters limit suitable landing spots for standard drones, but where there may still be vertical surfaces intact. Engineers at Stanford University have developed Scamp (pictured), a flying robot which can land on walls and climb vertical surfaces using its spindly legs. Called the Stanford Climbing and Aerial Manoeuvring Platform, affectionately shortened to Scamp, the flying robot is just as happy climbing surfaces as it is in the air. It uses two spindly, daddy long legs-style limbs to pull itself up walls and surfaces, and was developed and built at Stanford University's Biomimetics and Dextrous Manipulation Lab.