Progressive Fusion for Unsupervised Binocular Depth Estimation Using Cycled Networks.

Recent deep monocular depth estimation approaches based on supervised regression have achieved remarkable performance. However, they require costly ground truth annotations during training. To cope with this issue, in this paper we present a novel unsupervised deep learning approach for predicting depth maps. We introduce a new network architecture, named Progressive Fusion Network (PFN), that is specifically designed for binocular stereo depth estimation. This network is based on a multi-scale refinement strategy that combines the information provided by both stereo views. In addition, we propose to stack twice this network in order to form a cycle. This cycle approach can be interpreted as a form of data-augmentation since, at training time, the network learns both from the training set images (in the forward half-cycle) but also from the synthesized images (in the backward half-cycle). The architecture is jointly trained with adversarial learning. Extensive experiments on the publicly available datasets KITTI, Cityscapes and ApolloScape demonstrate the effectiveness of the proposed model which is competitive with other unsupervised deep learning methods for depth prediction.

Genetic diversity in widespread species is not congruent with species richness in alpine plant communities.

The Convention on Biological Diversity (CBD) aims at the conservation of all three levels of biodiversity, that is, ecosystems, species and genes. Genetic diversity represents evolutionary potential and is important for ecosystem functioning. Unfortunately, genetic diversity in natural populations is hardly considered in conservation strategies because it is difficult to measure and has been hypothesised to co-vary with species richness. This means that species richness is taken as a surrogate of genetic diversity in conservation planning, though their relationship has not been properly evaluated. We tested whether the genetic and species levels of biodiversity co-vary, using a large-scale and multi-species approach. We chose the high-mountain flora of the Alps and the Carpathians as study systems and demonstrate that species richness and genetic diversity are not correlated. Species richness thus cannot act as a surrogate for genetic diversity. Our results have important consequences for implementing the CBD when designing conservation strategies.

Recent plant diversity changes on Europe's mountain summits.

In mountainous regions, climate warming is expected to shift species' ranges to higher altitudes. Evidence for such shifts is still mostly from revisitations of historical sites. We present recent (2001 to 2008) changes in vascular plant species richness observed in a standardized monitoring network across Europe's major mountain ranges. Species have moved upslope on average. However, these shifts had opposite effects on the summit floras' species richness in boreal-temperate mountain regions (+3.9 species on average) and Mediterranean mountain regions (-1.4 species), probably because recent climatic trends have decreased the availability of water in the European south. Because Mediterranean mountains are particularly rich in endemic species, a continuation of these trends might shrink the European mountain flora, despite an average increase in summit species richness across the region.