Detection of small-sized targets in aerial views is a challenging task due to the smallness of vehicle size, complex background, and monotonic object appearances. In this letter, we propose a one-stage vehicle detection network (AVDNet) to robustly detect small-sized vehicles in aerial scenes. In AVDNet, we introduced ConvRes residual blocks at multiple scales to alleviate the problem of vanishing features for smaller objects caused because of the inclusion of deeper convolutional layers. These residual blocks, along with enlarged output feature map, ensure the robust representation of the salient features for smallsized objects. Furthermore, we proposed a recurrent-feature aware visualization (RFAV) technique to analyze the network behavior. We also created a new airborne image data set (ABD) by annotating 1396 new objects in 79 aerial images for our experiments. The effectiveness of AVDNet is validated on VEDAI, DLR-3K, DOTA, and the combined (VEDAI, DLR-3K, DOTA, and ABD) data set. Experimental results demonstrate the significant performance improvement of the proposed method over state-ofthe-art detection techniques in terms of mAP, computation, and space complexity.

No abstract

Purpose: To analyze the weekly rate of retinal vascular growth in treatment-naïve babies with various stages of retinopathy of prematurity (ROP) and validate if this could be a predictor of treatment need. Methods: Retrospective review of medical charts and retinal images of babies with various stages of ROP. The images were enhanced using red-green image enhancement software. Using the length of the horizontal disc diameter (DD) of each eye, the vessel growth was measured from the disc margin up to the vessel tip in fixed quadrants. The rate of vessel growth was the ratio of vessel length to the number of weeks it took to reach this length. The babies were divided into treatment warranting ROP (group 1), low-risk pre-threshold (type II) ROP (group 2,), and no-ROP (group 3) for analysis. The “no-ROP” group acted as normal control. Group 1 was further subdivided into 1A (threshold ROP), IB (aggressive posterior ROP), 1C (hybrid ROP), and ID (high-risk pre-threshold ROP). Results: Out of 436 eyes, groups 1, 2, and 3 had 238, 108, and 90 eyes, respectively. The mean rate of vascular outgrowth along with 95% confidence interval (CI) was 0.490 [0.487,0.520], 0.612 [0.599, 0.638], and 0.719 [0.703, 0.740] DD/week, respectively, for babies with “treatment warranting,” “low risk pre-threshold” and “no ROP” groups, respectively. In our estimate, more than 80% of eyes with a vessel growth rate of 0.54 DD/week or less required treatment Conclusion: A rate of retinal vascular growth less than 0.54 DD/week can be used to determine treatment requirements in babies with ROP.

In this paper, a stage-structured predator–prey model is proposed and analyzed with density-dependent maturation delay. We studied the dynamics of our model analytically and obtained conditions which influence the positivity and boundedness of all populations. Criteria for the existence of a non-trivial equilibrium and conditions for the uniqueness of this equilibrium are given. A linearized analysis on the equilibria, which is algebraically very complicated in the case of non-trivial equilibrium, is carried out. We proved that the system is globally asymptotically stable in the situation when non-trivial equilibrium does not exist. To accomplish our all analytical findings and to investigate the effect of density-dependent maturation delay on the system behavior, we presented a numerical simulation. It is concluded that variations in parameter, which we introduce in the system to observe the effect of density-dependent maturation delay, produces significant quantitative changes in system behavior and also qualitative changes in the behavior of immature predator population.

This paper deals with a ratio-dependent predator-prey model where the prey population is stage-structured consisting of immature and mature stages and the predator population is influenced by the resource biomass. By means of a transformation of variables, we transform the model into a dynamical system in such a way that there is one-to-one correspondence between the positive values of the original model and the positive values of the transformed model, so that the results which are true for the transformed model are also true for the original model. Dynamical behaviors such as positivity, boundedness, stability, bifurcation and persistence of the model are studied analytically using theory of differential equations. Computer simulations are carried out to substantiate the analytical findings. It is noted that the influence of the resource biomass on the predator population may lead to the extinction of prey population at a lesser value of maturity time in comparison to the absence of the resource biomass.