Existing enhancement methods are empirically expected to help the high-level end computer vision task: however, that is observed to not always be the case in practice. We focus on object or face detection in poor visibility enhancements caused by bad weathers (haze, rain) and low light conditions. To provide a more thorough examination and fair comparison, we introduce three benchmark sets collected in real-world hazy, rainy, and lowlight conditions, respectively, with annotated objects/faces. We launched the UG 2+ challenge Track 2 competition in IEEE CVPR 2019, aiming to evoke a comprehensive discussion and exploration about whether and how low-level vision techniques can benefit the high-level automatic visual recognition in various scenarios. To our best knowledge, this is the first and currently largest effort of its kind. Baseline results by cascading existing enhancement and detection models are reported, indicating the highly challenging nature of our new data as well as the large room for further technical innovations. Thanks to a large participation from the research community, we are able to analyze representative team solutions, striving to better identify the strengths and limitations of existing mindsets as well as the future directions.Index Terms-Poor visibility environment, object detection, face detection, haze, rain, low-light conditions *The first two authors Wenhan Yang and Ye Yuan contributed equally. Ye Yuan and Wenhan Yang helped prepare the dataset proposed for the UG2+ Challenges, and were the main responsible members for UG2+ Challenge 2019 (Track 2) platform setup and technical support. Wenqi Ren, Jiaying Liu, Walter J. Scheirer, and Zhangyang Wang were the main organizers of the challenge and helped prepare the dataset, raise sponsors, set up evaluation environment, and improve the technical submission. Other authors are the group members of winner teams in UG2+ challenge Track 2 contributing to the winning methods.

Our objective was to investigate the genetic polymorphisms of the glutathione S-transferase M1, T1, and P1 genes (GSTM1, GSTT1, and GSTP1) and to assess the oxidative damage in infertile men with varicoceles from northwestern China. A total of 65 infertile men with varicoceles and 30 controls were included in the study. Multiplex polymerase chain reaction and polymerase chain reaction-restriction fragment length polymorphism analyses were used to identify the genotypes. Sperm DNA damage was assessed by terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end-labeling (TUNEL). The levels of 8-hydroxy-29-deoxyguanosine (8-OH-dG) were measured by high-performance liquid chromatography with electrochemical detection. The activities of malondialdehyde (MDA) and nitric oxide (NO), and the total antioxidant capacity (TAC) were detected by spectroscopic analysis, and sperm characteristics were measured using computer-assisted semen analysis. The frequencies of the GSTM1, GSTT1, and GSTP1 genotypes were not significantly different between the control and patient groups (P . .05). The percentage of TUNEL-positive sperm and the levels of 8-OH-dG, MDA, and NO were higher but the sperm concentration and motility and the TAC were lower in the patients with the GSTM1, GSTT1, and GSTM1/T1 null genotypes than those in the patients with the GSTM1, GSTT1, and GSTM1/T1 present genotypes (P , .05). However, no significant differences were observed between the GSTP1 A/A and A/G+G/G genotypes (P . .05). Our results suggest that the GSTM1 and GSTT1 null genotypes may predispose sperm to increased oxidative damage in infertile men with varicoceles; however, GSTP1 allelic variation was not significantly different between the patient and control groups in this study.

Although cooperative social networks are considered key to human evolution, emphasis has usually been placed on the functions of men’s cooperative networks. What do women’s networks look like? Do they differ from men’s networks and what does this suggest about evolutionarily inherited gender differences in reproductive and social strategies? In this paper, we test the ‘universal gender differences’ hypothesis positing gender-specific network structures against the ‘gender reversal’ hypothesis that posits that women’s networks look more ‘masculine’ under matriliny. Specifically, we ask whether men’s friendship networks are always larger than women’s networks and we investigate measures of centrality by gender and descent system. To do so, we use tools from social network analysis and data on men’s and women’s friendship ties in matrilineal and patrilineal Mosuo communities. In tentative support of the gender reversal hypothesis, we find that women’s friendship networks in matriliny are relatively large. Measures of centrality and generalized linear models otherwise reveal greater differences between communities than between men and women. The data and analyses we present are primarily descriptive given limitations of sample size and sampling strategy. Nonetheless, our results provide support for the flexible application of social relationships across genders and clearly challenge the predominant narrative of universal gender differences across space and time.

IntroductionAdults can represent numerical information in nonsymbolic and symbolic formats and flexibly switch between the two. While some studies suggest a strong link between the two number representation systems (e.g., Piazza, Izard, Pinel, Le Bihan, & Dehaene, 2004 Neuron, 44(3), 547), other studies show evidence against the strong‐link hypothesis (e.g., Lyons, Ansari, & Beilock, 2012 Journal of Experimental Psychology: General, 141(4), 635). This inconsistency could arise from the relation between task demands and the closeness of the link between the two number systems.MethodsWe used a passive viewing task and event‐related potentials (ERP) to examine the temporal dynamics of the implicit integration between the nonsymbolic and symbolic systems. We focused on two ERP components over posterior scalp sites that were found to be sensitive to numerical distances and ratio differences in both numerical formats: a negative component that peaks around 170 ms poststimulus (N1) and a positive component that peaks around 200 ms poststimulus (P2p). We examined adults' (n = 55) ERPs when they were passively viewing simultaneously presented dot quantities and Arabic numerals (i.e., nonsymbolic and symbolic numerical information) in the double‐digit range. For each stimulus, the nonsymbolic and symbolic content either matched or mismatched in number. We also asked each participant to estimate dot quantities in a separate behavioral task and observed that they tended to underestimate the actual dot quantities, suggesting a need to adjust the match between nonsymbolic and symbolic information to reflect the perceived quantity of the nonsymbolic information.ResultsUsing this adjustment, participants showed greater N1 and P2p amplitudes when perceived dot quantities matched Arabic numerals than when there was a mismatch. However, no differences were found between the unadjusted match and mismatch conditions.ConclusionOur findings suggest that adults rapidly integrate nonsymbolic and symbolic formats of double‐digit numbers, but evidence of such integration is best observed when the perceived (rather than veridical) dot quantity is considered.

No abstract