MAU-Net: A Retinal Vessels Segmentation Method.

Detailed extraction of retinal vessel morphology is of great significance in many clinical applications. In this paper, we propose a retinal image segmentation method, called MAU-Net, which is based on the U-net structure and takes advantages of both modulated deformable convolution and dual attention modules to realize vessels segmentation. Specifically, based on the classic U-shaped architecture, our network introduces the Modulated Deformable Convolutional (MDC) block as encoding and decoding unit to model vessels with various shapes and deformations. In addition, in order to obtain better feature presentations, we aggregate the outputs of dual attention modules: the position attention module (PAM) and channel attention module (CAM). On three publicly available datasets: DRIVE, STARE and CHASEDB1, we have achieved superior performance to other algorithms. Quantitative and qualitative experimental results show that our MAU-Net can effectively and accurately accomplish the retinal vessels segmentation task.

Roosting Site Usage, Gregarious Roosting and Behavioral Interactions During Roost-assembly of Two Lycaenidae Butterflies.

Lycaenidae is one of the larger of the world's butterfly families, based on number and diversity of species, but knowledge of roosting in this group is sparse. Zizina otis riukuensis and Zizeeria maha okinawana are two small lycaenids that are commonly found in urban settings and widely distributed across much of Asia. We conducted experiments on a university campus to determine the plant species and plant structures commonly used by these two blues when roosting. We also tested the hypothesis that gregarious roosting exists in these two blues by demonstrating the non-random distribution of roosting blues and the tight mapping of their roosts to the spatial distribution of specific plant species and/or specific plant structures, as well as by demonstrating behavioral interactions among individuals during roosting-assembly. We found that both Z. otis and Z. maha roosted primarily on flowers and fruits of Tridax procumbens and Vernonia cinerea. We also found that these blues formed conspicuous roosting aggregations with significant positive associations between the flowers and fruits of both T. procumbens and V. cinerea and the blues. Moreover, our behavioral observations showed that these blues expressed various levels of interaction during roosting gatherings. Based on these findings, we conclude that gregarious roosting exists in both Z. otis and Z. maha. To our knowledge, this paper represents one of the first demonstration of nocturnal gregarious roosting in lycaenids. This study also highlights the importance of institutional estates in providing roosting resources for butterflies in urban ecosystems.