This paper extends the popular task of multi-object tracking to multi-object tracking and segmentation (MOTS). Towards this goal, we create dense pixel-level annotations for two existing tracking datasets using a semi-automatic annotation procedure. Our new annotations comprise 65,213 pixel masks for 977 distinct objects (cars and pedestrians) in 10,870 video frames. For evaluation, we extend existing multi-object tracking metrics to this new task. Moreover, we propose a new baseline method which jointly addresses detection, tracking, and segmentation with a single convolutional network. We demonstrate the value of our datasets by achieving improvements in performance when training on MOTS annotations. We believe that our datasets, metrics and baseline will become a valuable resource towards developing multi-object tracking approaches that go beyond 2D bounding boxes. We make our annotations, code, and models available at https: //www.vision.rwth-aachen.de/page/mots.
The deposition of Mn onto Si͑001͒ in the submonolayer regime has been studied with scanning tunneling microscopy to gain insight into the bonding and energetics of Mn with Si. The as-deposited Mn films at room temperature are unstructured. Upon annealing to 300-700°C three-dimensional islands of Mn or Mn x Si y form while between the islands the Si͑001͒-͑2 ϫ 1͒ reconstruction becomes visible. With increasing annealing time the density of islands per surface area decreases while the average height of the remaining islands increases. The large islands grow in size at the expense of the small ones, which can be understood in the context of Ostwald ͓Z. Phys. Chem. 34, 495 ͑1900͔͒ ripening theory. The average island height shows a time dependence of H ϳ t 1/4 , indicating that surface diffusion is the growth limiting process.
The deposition of Mn atoms onto the Si͑001͒-͑2 ϫ 1͒ reconstructed surface has been studied using scanning tunneling microscopy ͑STM͒ and first-principles electronic structure calculations. Room-temperature deposition of 0.1 ML ͑monolayer͒ of Mn gives rise to a disordered surface structure. After in situ annealing between 300 and 700°C, most of the Mn is incorporated into three-dimensional manganese silicide islands, and Si dimer rows reappear in the STM images on most of the substrate surface. At the same time, rowlike structures are visible in the atomic-scale STM images. A comparison with calculated STM images provides evidence that Mn atoms are incorporated into the row structures in subsurface interstitial sites, which are the lowest-energy position for Mn on Si͑001͒. The subsurface Mn alters the height and local density of states of the Si dimer atoms, causing them to appear 0.6 Å higher than a neighboring Si dimer with no Mn below. This height difference that allows the detection the subsurface Mn results from a subtle interplay of geometrical and electronic effects.
Articles you may be interested inUltrahigh vacuum instrument that combines variable-temperature scanning tunneling microscopy with Fourier transform infrared reflection-absorption spectroscopy for studies of chemical reactions at surfaces A molecular beam epitaxy and low temperature scanning tunneling microscopy chamber have been integrated to characterize both compound and elemental semiconductor surfaces and interfaces. The integration of these two commercially available systems has been achieved using a custom designed sample transfer mechanism. The MBE growth chamber is equipped with electron diffraction and provides substrate temperature measurements and control by means of band-edge thermometry accurate to within ±0.5°C. In addition, the microscope can operate at temperatures as low as 4 K and perform ballistic electron emission microscopy measurements. The chamber that houses the microscope includes a preparation chamber with an evaporation source for metals. The entire STM chamber also rests on an active vibration isolation table, while still maintaining an all ultrahigh vacuum connection to the MBE system.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.