Reconstruction Guided Meta-Learning for Few Shot Open Set Recognition.

In many applications, we are constrained to learn classifiers from very limited data (few-shot classification). The task becomes even more challenging if it is also required to identify samples from unknown categories (open-set classification). Learning a good abstraction for a class with very few samples is extremely difficult, especially under open-set settings. As a result, open-set recognition has received limited attention in the few-shot setting. However, it is a critical task in many applications like environmental monitoring, where the number of labeled examples for each class is limited. Existing few-shot open-set recognition (FSOSR) methods rely on thresholding schemes, with some considering uniform probability for open-class samples. However, this approach is often inaccurate, especially for fine-grained categorization, and makes them highly sensitive to the choice of a threshold. To address these concerns, we propose Reconstructing Exemplar-based Few-shot Open-set ClaSsifier (ReFOCS). By using a novel exemplar reconstruction-based meta-learning strategy ReFOCS streamlines FSOSR eliminating the need for a carefully tuned threshold by learning to be self-aware of the openness of a sample. The exemplars, act as class representatives and can be either provided in the training dataset or estimated in the feature domain. By testing on a wide variety of datasets, we show ReFOCS to outperform multiple state-of-the-art methods.

Learning Person Re-Identification Models From Videos With Weak Supervision.

Most person re-identification methods, being supervised techniques, suffer from the burden of massive annotation requirement. Unsupervised methods overcome this need for labeled data, but perform poorly compared to the supervised alternatives. In order to cope with this issue, we introduce the problem of learning person re-identification models from videos with weak supervision. The weak nature of the supervision arises from the requirement of video-level labels, i.e. person identities who appear in the video, in contrast to the more precise frame-level annotations. Towards this goal, we propose a multiple instance attention learning framework for person re-identification using such video-level labels. Specifically, we first cast the video person re-identification task into a multiple instance learning setting, in which person images in a video are collected into a bag. The relations between videos with similar labels can be utilized to identify persons, on top of that, we introduce a co-person attention mechanism which mines the similarity correlations between videos with person identities in common. The attention weights are obtained based on all person images instead of person tracklets in a video, making our learned model less affected by noisy annotations. Extensive experiments demonstrate the superiority of the proposed method over the related methods on two weakly labeled person re-identification datasets.

Context-Aware Query Selection for Active Learning in Event Recognition.

Activity recognition is a challenging problem with many practical applications. In addition to the visual features, recent approaches have benefited from the use of context, e.g., inter-relationships among the activities and objects. However, these approaches require data to be labeled, entirely available beforehand, and not designed to be updated continuously, which make them unsuitable for surveillance applications. In contrast, we propose a continuous-learning framework for context-aware activity recognition from unlabeled video, which has two distinct advantages over existing methods. First, it employs a novel active-learning technique that not only exploits the informativeness of the individual activities but also utilizes their contextual information during query selection; this leads to significant reduction in expensive manual annotation effort. Second, the learned models can be adapted online as more data is available. We formulate a conditional random field model that encodes the context and devise an information-theoretic approach that utilizes entropy and mutual information of the nodes to compute the set of most informative queries, which are labeled by a human. These labels are combined with graphical inference techniques for incremental updates. We provide a theoretical formulation of the active learning framework with an analytic solution. Experiments on six challenging datasets demonstrate that our framework achieves superior performance with significantly less manual labeling.

Exploiting Typicality for Selecting Informative and Anomalous Samples in Videos.

In this paper, we present a novel approach to find informative and anomalous samples in videos exploiting the concept of typicality from information theory. In most video analysis tasks, selection of the most informative samples from a huge pool of training data in order to learn a good recognition model is an important problem. Furthermore, it is also useful to reduce the annotation cost as it is time-consuming to annotate unlabeled samples. Typicality is a simple and powerful technique which can be applied to compress the training data to learn a good classification model. In a continuous video clip, an activity shares a strong correlation with its previous activities. We assume that the activity samples that appear in a video form a Markov chain. We explicitly show how typicality can be utilized in this scenario. We compute an atypical score for a sample using typicality and the Markovian property, which can be applied to two challenging vision problems-(a) sample selection for learning activity recognition models, and (b) anomaly detection. In the first case, our approach leads to a significant reduction of manual labeling cost while achieving similar or better recognition performance compared to a model trained with the entire training set. For the latter case, the atypical score has been exploited in identifying anomalous activities in videos where our results demonstrate the effectiveness of the proposed framework over other recent strategies.

Simultaneous wavelength and orbital angular momentum demultiplexing using tunable MEMS-based Fabry-Perot filter.

In this paper, we experimentally demonstrate simultaneous wavelength and orbital angular momentum (OAM) multiplexing/demultiplexing of 10 Gbit/s data streams using a new on-chip micro-component-tunable MEMS-based Fabry-Perot filter integrated with a spiral phase plate. In the experiment, two wavelengths, each of them carrying two channels with zero and nonzero OAMs, form four independent information channels. In case of spacing between wavelength channels of 0.8 nm and intensity modulation, power penalties relative to the transmission of one channel do not exceed 1.45, 0.79 and 0.46 dB at the hard-decision forward-error correction (HD-FEC) bit-error-rate (BER) limit 3.8 × 10-3 when multiplexing a Gaussian beam and OAM beams of azimuthal orders 1, 2 and 3 respectively. In case of phase modulation, power penalties do not exceed 1.77, 0.54 and 0.79 dB respectively. At the 0.4 nm wavelength grid, maximum power penalties at the HD-FEC BER threshold relative to the 0.8 nm wavelength spacing read 0.83, 0.84 and 1.15 dB when multiplexing a Gaussian beam and OAM beams of 1st, 2nd and 3rd orders respectively. The novelty and impact of the proposed filter design is in providing practical, integrable, cheap, and reliable transformation of OAM states simultaneously with the selection of a particular wavelength in wavelength division multiplexing (WDM). The proposed on-chip device can be useful in future high-capacity optical communications with spatial- and wavelength-division multiplexing, especially for short-range communication links and optical interconnects.

Far-field, linewidth and thermal characteristics of a high-speed 1550-nm MEMS tunable VCSEL.

We report an electrically pumped 1550 nm MEMS tunable VCSEL with a continuous tuning of 101 nm at 22 °C. The top MEMS-DBR with built-in stress gradient within the dielectric layers is deposited in a low-temperature PECVD chamber on an InP-based half-VCSEL, structured by surface-micromachining and electrothermally actuated for continuous wavelength tuning. With 2.6 mA threshold current, the laser shows maximum CW output power of 3.2 mW at 1560 nm. The MEMS-VCSEL operates in single-mode with SMSR > 39 dB across the entire tuning range. At 36 °C, the tuning range reaches up to 107 nm. The divergence angle of the MEMS-VCSEL is approximately 5.6° for all tuning wavelengths. The intrinsic linewidth of an unpackaged device is 21 MHz. Quasi-error-free operation at 12.5 Gbps using a directly modulated MEMS-VCSEL is reported for a record 60 nm tuning, showing the potential of the so-called colorless source in WDM applications.

Wavelength-selective orbital-angular-momentum beam generation using MEMS tunable Fabry-Perot filter.

We demonstrate an on-chip device capable of wavelength-selective generation of vortex beams, which is realized by a spiral phase plate integrated onto a microelectromechanical system (MEMS) tunable filter. This vortex MEMS filter, being capable of functioning simultaneously in both wavelength and orbital-angular-momentum (OAM) domains at the 1550 nm wavelength regime, is considered as a compact, robust, and cost-effective solution for simultaneous OAM- and wavelength-division multiplexed optical communications. The experimental OAM spectra for azimuthal orders 1, 2, and 3 show an OAM state purity >92% across a wavelength range of more than 30 nm.

Widely tunable telecom MEMS-VCSEL for terahertz photomixing.

We report frequency-tunable terahertz (THz) generation with a photomixer driven by an ultra-broadband tunable micro-electro-mechanical system vertical-cavity surface-emitting laser (MEMS-VCSEL) and a fixed-wavelength VCSEL, as well as a tunable MEMS-VCSEL mixed with a distributed feedback (DFB) diode. A total frequency span of 3.4 THz is covered in direct detection mode and 3.23 THz in the homodyne mode. The tuning range is solely limited by the dynamic range of the photomixers and the Schottky diode/photoconductor used in the experiment.

Endovascular management in abdominal visceral arterial aneurysms: A pictorial essay.

Visceral artery aneurysms (VAAs) include aneurysms of the splanchnic circulation and those of the renal artery. Their diagnosis is clinically important because of the associated high mortality and potential complications. Splenic, superior mesenteric, gastroduodenal, hepatic and renal arteries are some of the common arteries affected by VAAs. Though surgical resection and anastomosis still remains the treatment of choice in some of the cases, especially cases involving the proximal arteries, increasingly endovascular treatment is being used for more distal vessels. We present a pictorial review of various intra-abdominal VAAs and their endovascular management.