• ### Deep Region Hashing for Efficient Large-scale Instance Search from Images(1701.07901)

Jan. 26, 2017 cs.CV
Instance Search (INS) is a fundamental problem for many applications, while it is more challenging comparing to traditional image search since the relevancy is defined at the instance level. Existing works have demonstrated the success of many complex ensemble systems that are typically conducted by firstly generating object proposals, and then extracting handcrafted and/or CNN features of each proposal for matching. However, object bounding box proposals and feature extraction are often conducted in two separated steps, thus the effectiveness of these methods collapses. Also, due to the large amount of generated proposals, matching speed becomes the bottleneck that limits its application to large-scale datasets. To tackle these issues, in this paper we propose an effective and efficient Deep Region Hashing (DRH) approach for large-scale INS using an image patch as the query. Specifically, DRH is an end-to-end deep neural network which consists of object proposal, feature extraction, and hash code generation. DRH shares full-image convolutional feature map with the region proposal network, thus enabling nearly cost-free region proposals. Also, each high-dimensional, real-valued region features are mapped onto a low-dimensional, compact binary codes for the efficient object region level matching on large-scale dataset. Experimental results on four datasets show that our DRH can achieve even better performance than the state-of-the-arts in terms of MAP, while the efficiency is improved by nearly 100 times.
• ### Bidirectional Long-Short Term Memory for Video Description(1606.04631)

June 15, 2016 cs.CL, cs.MM
Video captioning has been attracting broad research attention in multimedia community. However, most existing approaches either ignore temporal information among video frames or just employ local contextual temporal knowledge. In this work, we propose a novel video captioning framework, termed as \emph{Bidirectional Long-Short Term Memory} (BiLSTM), which deeply captures bidirectional global temporal structure in video. Specifically, we first devise a joint visual modelling approach to encode video data by combining a forward LSTM pass, a backward LSTM pass, together with visual features from Convolutional Neural Networks (CNNs). Then, we inject the derived video representation into the subsequent language model for initialization. The benefits are in two folds: 1) comprehensively preserving sequential and visual information; and 2) adaptively learning dense visual features and sparse semantic representations for videos and sentences, respectively. We verify the effectiveness of our proposed video captioning framework on a commonly-used benchmark, i.e., Microsoft Video Description (MSVD) corpus, and the experimental results demonstrate that the superiority of the proposed approach as compared to several state-of-the-art methods.
• ### Direct mass measurements of neutron-rich $^{86}$Kr projectile fragments and the persistence of neutron magic number $N$ = 32 in Sc isotopes(1603.06404)

March 21, 2016 nucl-ex
In this paper, we present direct mass measurements of neutron-rich $^{86}$Kr projectile fragments conducted at the HIRFL-CSR facility in Lanzhou by employing the Isochronous Mass Spectrometry (IMS) method. The new mass excesses of $^{52-54}$Sc nuclides are determined to be -40492(82), -38928(114), -34654(540) keV, which show a significant increase of binding energy compared to the reported ones in the Atomic Mass Evaluation 2012 (AME12). In particular, $^{53}$Sc and $^{54}$Sc are more bound by 0.8 MeV and 1.0 MeV, respectively. The behavior of the two neutron separation energy with neutron numbers indicates a strong sub-shell closure at neutron number $N$ = 32 in Sc isotopes.
• ### A data analysis method for isochronous mass spectrometry using two time-of-flight detectors at CSRe(1603.07169)

The concept of isochronous mass spectrometry (IMS) applying two time-of-flight (TOF) detectors originated many years ago at GSI. However, the corresponding method for data analysis has never been discussed in detail. Recently, two TOF detectors have been installed at CSRe and the new working mode of the ring is under test. In this paper, a data analysis method for this mode is introduced and tested with a series of simulations. The results show that the new IMS method can significantly improve mass resolving power via the additional velocity information of stored ions. This improvement is especially important for nuclides with Lorentz factor $\gamma$-value far away from the transition point $\gamma _t$ of the storage ring CSRe.