Accurate and robust detection of multi-class objects in very high resolution (VHR) aerial images has been playing a significant role in many real-world applications. The traditional detection methods have made remarkable progresses with horizontal bounding boxes (HBBs) due to CNNs. However, HBB detection methods still exhibit limitations including the missed detection and the redundant detection regions, especially for densely-distributed and strip-like objects. Besides, large scale variations and diverse background also bring in many challenges. Aiming to address these problems, an effective region-based object detection framework named Multi-scale Feature Integration Attention Rotation Network (MFIAR-Net) is proposed for aerial images with oriented bounding boxes (OBBs), which promotes the integration of the inherent multi-scale pyramid features to generate a discriminative feature map. Meanwhile, the double-path feature attention network supervised by the mask information of ground truth is introduced to guide the network to focus on object regions and suppress the irrelevant noise. To boost the rotation regression and classification performance, we present a robust Rotation Detection Network, which can generate efficient OBB representation. Extensive experiments and comprehensive evaluations on two publicly available datasets demonstrate the effectiveness of the proposed framework.
With the help of a magnetic flux variable, the effects of stochastic electromagnetic disturbances on autapse Hodgkin-Huxley neuronal systems are studied systematically. Firstly, owing to the autaptic function, the inter-spike interval series of an autapse neuron not only bifurcates, but also presents a quasi-periodic characteristic. Secondly, an irregular mixed-mode oscillation induced by a specific electromagnetic disturbance is analyzed using the coefficient of variation of inter-spike intervals. It is shown that the neuronal discharge activity has certain selectivity to the noise intensity, and the appropriate noise intensity can induce the significant mixed-mode oscillations. Finally, the modulation effects of electromagnetic disturbances on a ring field-coupled neuronal network with autaptic structures are explored quantitatively using the average spiking frequency and the average coefficient of variation. The electromagnetic disturbances can not only destroy the continuous and synchronous discharge state, but also induce the resting neurons to generate the intermittent discharge mode and realize the transmission of neural signals in the neuronal network. The studies can provide some theoretical guidance for applying electromagnetic disturbances to effectively control the propagation of neural signals and treat mental illness.
The autaptic structure of neurons has the function of self-feedback, which is easily disturbed due to the quantum characteristics of neurotransmitter release. This paper focuses on the effect of conductance disturbance of chemical autapse on the electrophysiological activities of FHN neuron. First, the frequency encoding of FHN neuron to periodic excitation signals exhibits a nonlinear change characteristic, and the FHN neuron without autapse has chaotic discharge behavior according to the maximum Lyapunov exponent and the sampled time series. Secondly, the chemical autaptic function can change the dynamics of FHN neuronal system, and appropriate autaptic parameters can cause the dynamic bifurcation, which corresponds to the transition between different periodic spiking modes. In particular, the self-feedback function of chemical autapse can induce a transition from a chaotic discharge state to a periodic spiking or a quasi-periodic bursting discharge state. Finally, based on the quantum characteristics of neurotransmitter release, the effect of random disturbance from autaptic conductance on the firing activities is quantitatively studied with the help of the discharge frequency and the coefficient of variation of inter-spike interval series. The numerical results show that the disturbance of autaptic conductance can change the activity of ion channels under the action of self-feedback, which not only improves the encoding efficiency of FHN neuron to external excitation signals, but also changes the regularity of neuronal firing activities and induces significant coherent or stochastic bi-resonance. The coherent or stochastic bi-resonance phenomenon is closely related to the dynamic bifurcation of FitzHugh-Nagumo(FHN) neuronal system, and its underlying mechanism is that the disturbance of autaptic conductance leads to the unstable dynamic behavior of neuronal system, and the corresponding neuronal firing activity may transit between the resting state, the single-cycle and the multicycle spike states, thereby providing the occurring possibility for coherent or stochastic bi-resonance. This study further reveals the self-regulatory effect of the autaptic structure on neuronal firing activities, and could provide theoretical guidance for physiological manipulation of autapses. In addition, according to the pronounced self-feedback function of autaptic structure, a recurrent spiking neural network with local self-feedback can be constructed to improve the performance of machine learning by applying a synaptic plasticity rule.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.