Search

Dynamic scene understanding is one of the most conspicuous field of interest among computer vision community. In order to enhance dynamic scene understanding, pixel-wise segmentation with neural networks is widely accepted. The latest researches on pixel-wise segmentation combined semantic and motion information and produced good performance. In this work, we propose a state of art architecture of neural networks to accurately and efficiently get the moving object proposals (MOP). We first train an unsupervised convolutional neural network (UnFlow) to generate optical flow estimation. Then we render the output of optical flow net to a fully convolutional SegNet model. The main contribution of our work is (1) Fine-tuning the pretrained optical flow model on the brand new DAVIS Dataset; (2) Leveraging fully convolutional neural networks with Encoder-Decoder architecture to segment objects. We developed the codes with TensorFlow, and executed the training and evaluation processes on an AWS EC2 instance., Comment: 7 pages, 8 figures, 1 table

Development of advanced nanophotonic devices is currently in rapid growth and revolutionizing the whole fields of integrated optics and photonics. Devices such as nanoscale LEDs & lasers, waveguide couplers and modulators are essential components in applications ranging from light sources to optical circuits and quantum information processing. The optical characteristics of these nanostructures could be engineered to realize strong confinement of optical modes within their low dimensions, which leads to strong light-matter interactions at desirable wavelength range when coupling to high-efficient, low-dimensional quantum emitters such as colloidal nanoplatelets, perovskite nanocrystals and transition metal dichalcogenide monolayers with unique optical properties. These light-matter coupled systems could realize various kinds of nanophotonic devices with high efficiency and nonlinearity in development of more complex optical circuits and quantum networks. In this thesis, I present my work first on experimental demonstration of spontaneous emission intensity and rate enhancement of both colloidal cadmium selenide/cadmium sulfide core/shell nanoplatelet and cesium halide bromide perovskite nanocrystals in Purcell regime by using silicon nitride photonic crystal nanobeam cavities. The one-dimensional high-quality cavity confines the emission in a small mode volume with high radiative decay mode density, leading to a clear increase in their photoluminescence efficiencies. We next present realization of a continuous-wave nanolaser based on this coupled system operating at room temperature. The high coupling efficiency results in a record-low pump threshold at 1 μW. This result shows that colloidal nanocrystals are suitable for compact and efficient opto-electronic devices based on solution-processable materials. Besides light generation, furthermore for transmission and processing, we have also realized chiral light-matter interactions in a glide-plane photonic crystal wa

Metastatic follicular thyroid carcinoma (FTC), unresectable or resistance to radioactive iodine, is associated with poor survival. It is believed that this kind of FTC is driven by mutated genes. However, what kind of changes of genome and underlying mechanisms are elusive. The aim of this article is to understand whether there are somatic mutations in circulating cell-free tumor DNA (cfDNA) in a FTC patient with lung and bone metastases. A 55-year-old woman was diagnosed with FTC with bone and lung metastases. Appropriate amounts of DNA were extracted from formalin-fixed, paraffin-embedded thyroid tumor, peripheral cell-free plasma, and peripheral blood leukocytes and then sequenced. The significance of DNA sequencing was evaluated. There were 13,519 common variants in both tissue DNA and cfDNA. Fifty-five somatic mutations were identified in tumor, with 5 of them nonsynonymous. Seventy-two somatic mutations were found in cfDNA, with 2 of them causing amino acid change. Sixteen common alterations existed in both samples, that is, 31.3% of all the tissue somatic mutations. This pilot study provided proof that cfDNA represents the genomic characteristics of FTC primary tissue DNA well, but also metastatic tumors. Further studies are needed to better prove the effectiveness of cfDNA in the field of thyroid cancer metastatic mechanism research and real-time monitoring.

We propose a high-precision method for measuring the orbital angular momentum (OAM) spectrum of ultra-broadband optical-vortex (OV) pulses from fork-like interferograms between OV pulses and a reference plane-wave pulse. It is based on spatial reconstruction of the electric fields of the pulses to be measured from the frequency-resolved interference pattern. Our method is demonstrated experimentally by obtaining the OAM spectra for different spectral components of the OV pulses, enabling us to characterize the frequency dispersion of the topological charge of the OAM spectrum by a simple experimental setup. Retrieval is carried out in quasi-real time, allowing us to investigate OAM spectra dynamically. Furthermore, we determine the relative phases (including the sign) of the topological-charge-resolved electric-field amplitudes, which are significant for evaluating OVs or OV pulses with arbitrarily superposed modes.

We propose a high-precision method for measuring the orbital angular momentum (OAM) spectrum of ultra-broadband optical-vortex (OV) pulses from fork-like interferograms between OV pulses and a reference plane-wave pulse. It is based on spatial reconstruction of the electric fields of the pulses to be measured from the frequency-resolved interference pattern. Our method is demonstrated experimentally by obtaining the OAM spectra for different spectral components of the OV pulses, enabling us to characterize the frequency dispersion of the topological charge of the OAM spectrum by a simple experimental setup. Retrieval is carried out in quasi-real time, allowing us to investigate OAM spectra dynamically. Furthermore, we determine the relative phases (including the sign) of the topological-charge-resolved electric-field amplitudes, which are significant for evaluating OVs or OV pulses with arbitrarily superposed modes.

This paper examines users who are common to two popular online social networks, Instagram and Ask.fm, that are often used for cyberbullying. An analysis of the negativity and positivity of word usage in posts by common users of these two social networks is performed. These results are normalized in comparison to a sample of typical users in both networks. We also examine the posting activity of common user profiles and consider its correlation with negativity. Within the Ask.fm social network, which allows anonymous posts, the relationship between anonymity and negativity is further explored.