In this paper, we first provide a new perspective to divide existing high performance object detection methods into direct and indirect regressions. Direct regression performs boundary regression by predicting the offsets from a given point, while indirect regression predicts the offsets from some bounding box proposals. Then we analyze the drawbacks of the indirect regression, which the recent state-of-the-art detection structures like Faster-RCNN and SSD follows, for multi-oriented scene text detection, and point out the potential superiority of direct regression. To verify this point of view, we propose a deep direct regression based method for multi-oriented scene text detection. Our detection framework is simple and effective with a fully convolutional network and one-step post processing. The fully convolutional network is optimized in an end-to-end way and has bi-task outputs where one is pixel-wise classification between text and non-text, and the other is direct regression to determine the vertex coordinates of quadrilateral text boundaries. The proposed method is particularly beneficial for localizing incidental scene texts. On the ICDAR2015 Incidental Scene Text benchmark, our method achieves the F1-measure of 81%, which is a new state-ofthe-art and significantly outperforms previous approaches. On other standard datasets with focused scene texts, our method also reaches the state-of-the-art performance.

No abstract

To address the problem of estimating camera trajectory and to build a structural three-dimensional (3D) map based on inertial measurements and visual observations, this paper proposes point–line visual–inertial odometry (PL-VIO), a tightly-coupled monocular visual–inertial odometry system exploiting both point and line features. Compared with point features, lines provide significantly more geometrical structure information on the environment. To obtain both computation simplicity and representational compactness of a 3D spatial line, Plücker coordinates and orthonormal representation for the line are employed. To tightly and efficiently fuse the information from inertial measurement units (IMUs) and visual sensors, we optimize the states by minimizing a cost function which combines the pre-integrated IMU error term together with the point and line re-projection error terms in a sliding window optimization framework. The experiments evaluated on public datasets demonstrate that the PL-VIO method that combines point and line features outperforms several state-of-the-art VIO systems which use point features only.

No abstract

The complexity of chemical compounds in lithium‐ion batteries (LIBs) results in great difficulties in the extraction of multiple transition metals, which have similar physicochemical characteristics. Here, we propose a novel strategy for selective extraction of nickel, cobalt, and manganese from spent LiNixCoyMn1−x−yO2 (NCM) cathode through the regulation of coordination environment. Depending on adjusting the composition of ligand in transition metal complexes, a tandem leaching and separation system is designed and finally enables nickel, cobalt, and manganese to enrich in the form of NiO, Co3O4, and Mn3O4 with high recovery yields of 99.1 %, 95.1 %, and 95.3 %, respectively. We further confirm that the combination of different transition metals with well‐designed ligands is the key to good selectivity. Through our work, fine‐tuning the coordination environment of metal ions is proved to have great prospects in the battery recycling industry.

Background: Recent guidelines provide better treatment and management of pregnancy in women with systemic lupus erythematosus (SLE). In this analysis, we aimed to systematically assess the maternal and fetal complications associated with SLE using the most recent studies (2017–2019) to obtain an updated result of the present situation. Methods: http://www.clinicaltrials.gov , MEDLINE, Cochrane Central, Web of Science, EMBASE, and Google Scholar were searched for English based studies comparing maternal and fetal complications in pregnant women with versus without SLE. Maternal and fetal complications were the endpoints in this analysis. The RevMan software 5.3 (latest version) was the most suitable analytical software for this analysis. Data were represented by risk ratio (RR) with 95% confidence interval (CI). Results: A total number of eight million eight hundred and twelve thousand two hundred seventy-two (8,812,272) participants were included in this analysis, consisting of 9696 SLE-associated pregnancy. Based on an analysis of recently published studies (2017–2019), pre-eclampsia/eclampsia was significantly higher in pregnant women with SLE (RR: 3.38, 95% CI: 3.15–3.62; P = .00001). SLE was also associated with an increased risk of stillbirth (RR: 16.49, 95% CI: 2.95–92.13; P = .001) and fetal loss (RR: 7.55, 95% CI: 4.75–11.99; P = .00001). Abortion (RR: 4.70, 95% CI: 3.02–7.29; P = .00001) and the risk for cesarean section due to complications (RR: 1.38, 95% CI: 1.11–1.70; P = .003) were also significantly higher in pregnant women with SLE. In addition, fetal complications including preterm birth (RR: 2.33, 95% CI: 1.78–3.05; P = .00001), infants who were small for gestational age (RR: 2.50, 95% CI: 1.41–4.45; P = .002) and infants with low birth weight (RR: 4.78, 95% CI: 3.65–6.26; P = .00001) were also significantly higher in newborns from mothers with SLE. Moreover, the risk of newborns who were admitted to the neonatal intensive care unit (RR: 2.79, 95% CI: 2.31–3.37; P = .00001), newborns with an APGAR score <7 within 1 minute (RR: 2.47, 95% CI: 1.68–3.62; P = .00001) and 5 minutes (RR: 3.63, 95% CI: 2.04–6.45; P = .0001) respectively, were significantly highly associated with SLE. Conclusions: Based on the most recent studies, we could conclude that maternal and fetal complications were significantly higher in SLE-associated pregnancy. Therefore, SLE should still be considered a severe risk factor for pregnancy.

As one of the fascinating high capacity cathodes, O3-type layered oxides usually suffer from their intrinsic air sensitivity and sluggish kinetics originating from the spontaneous lattice Na extraction during air exposure and high tetrahedral site energy of Na + diffusion transition state. What is worse, the improvement on the two handicaps is hard to simultaneously realize because of the contradiction between Na containment suggested in air stability mechanism and enhanced Na diffusion mentioned in kinetics strategy. Herein, it is shown that a simple strategy of introducing proper Na vacancies into lattice can simultaneously realize a dual performance improvement. Na vacancies decrease the charge density on transitional metal ions and enhance the antioxidative capability of material, ensuring a stable lattice Na containment for Na 0.93 Li 0.12 Ni 0.25 Fe 0.15 Mn 0.48 O 2 when exposed to air. Additionally, more Na + diffusional sites and enlarged Na layer spacing are obtained and result in a significantly decreased energy barrier from ≈1000 to 300 meV and a high rate capability of 70.8% retention at 2000 mA g −1 .Remarkably, such a strategy can be easily realized by either pre-or posttreating, which exhibits excellent universality for various O3 materials, implying its enormous potential to promote the commercial application of O3-type cathodes.

No abstract