Some time division multiple access (TDMA) wireless cooperative relay protocols have recently been developed. The works about their symbol error rate (SER) performance, however, are limited to SER upper bounds or SER f...
详细信息
This paper is to predict economic development based on wavelet least squares support vector machine algorithm. The same as support vector machine, least squares support vector machine employs the principle of structur...
详细信息
While promising results have been achieved in weakly-supervised semantic segmentation (WSSS), limited supervision from image-level tags inevitably induces discriminative reliance and spurious relations between target ...
详细信息
While promising results have been achieved in weakly-supervised semantic segmentation (WSSS), limited supervision from image-level tags inevitably induces discriminative reliance and spurious relations between target classes and background regions. Thus, Class Activation Map (CAM) usually tends to activate discriminative object regions and falsely includes lots of class-related backgrounds. Without pixel-level supervisions, it could be very difficult to enlarge the foreground activation and suppress those false activation of background regions. In this paper, we propose a novel framework of Cross Language Image Matching with Automatic Context Discovery (CLIMS++), based on the recently introduced Contrastive Language-Image Pre-training (CLIP) model, for WSSS. The core idea of our framework is to introduce natural language supervision to activate more complete object regions and suppress class-related background regions in CAM. In particular, we design object, background region, and text label matching losses to guide the model to excite more reasonable object regions of each category. In addition, we propose to automatically find spurious relations between foreground categories and backgrounds, through which a background suppression loss is designed to suppress the activation of class-related backgrounds. The above designs enable the proposed CLIMS++ to generate a more complete and compact activation map for the target objects. Extensive experiments on PASCAL VOC 2012 and MS COCO 2014 datasets show that our CLIMS++ significantly outperforms the previous state-of-the-art methods.
Some Time Division Multiple Access (TDMA) wireless cooperative relay protocols and their Symbol Error Rate (SER) performance have recently been developed. These works, however, are limited to SER upper bounds or SER f...
详细信息
Based on a Pade approximation, a wide-angle parabolic equation method is introduced for computing the multiobject radar cross section （RCS） for the first time. The method is a paraxial version of the scalar wave equ...
详细信息
Based on a Pade approximation, a wide-angle parabolic equation method is introduced for computing the multiobject radar cross section （RCS） for the first time. The method is a paraxial version of the scalar wave equation, which solves the field by marching them along the paraxial direction. Numerical results show that a single wide-angle parabofic equation run can compute multi-object RCS efficiently for angles up to 45 ° . The method provides anew and efficient numerical method for computation electromagnetics.
In tile process of the reconstruction of digital holography. the traditional methods of diffraction and filtration are commonly adopted to recover the original complex-wave signal. Influenced by twin-image and zero-or...
详细信息
In tile process of the reconstruction of digital holography. the traditional methods of diffraction and filtration are commonly adopted to recover the original complex-wave signal. Influenced by twin-image and zero-order terms, the above-mentioned methods, however, either limit tile field of vision or result in the loss of the amplitude and phase. A new method for complex-wave retrieval is presented, which is based on blind signal separation. Three frames of holograms are captured by a charge coupled device （CCD） camera to form an observation signal. The term containing only amplitude and phase of complex-wave is separated, by means of independent component analysis, from the observation signal, which effectively eliminates the zero-order term. Finally. the complex-wave retrieval of pure phase wavefront is achieved. Experimental results show that this method can better recover the amplitude and phase of the original complex-wave even when there is a frequency spectrum mixture in the hologram.
In this paper, a simplest fractional-order delayed memristive chaotic system is proposed in order to control the chaos behaviors via sliding mode control strategy. Firstly, we design a sliding mode control strategy fo...
详细信息
In this paper, a simplest fractional-order delayed memristive chaotic system is proposed in order to control the chaos behaviors via sliding mode control strategy. Firstly, we design a sliding mode control strategy for the fractionalorder system with time delay to make the states of the system asymptotically stable. Then, we obtain theoretical analysis results of the control method using Lyapunov stability theorem which guarantees the asymptotic stability of the noncommensurate order and commensurate order system with and without uncertainty and an external disturbance. Finally,numerical simulations are given to verify that the proposed sliding mode control method can eliminate chaos and stabilize the fractional-order delayed memristive system in a finite time.
In this paper, we focus on the Hopf bifurcation control of a small-world network model with time-delay. With emphasis on the relationship between the Hopf bifurcation and the time-delay, we investigate the effect of t...
详细信息
The noncontact detection methods of blood volume pulse (BVP) based on facial videos have become a hot spot in recent years. However, these kinds of methods are highly sensitive to face movement. To address this proble...
详细信息
To facilitate the integration of learning resources categorized under different ontology representations, the techniques of ontology mapping can be applied. Though many algorithms and systems have been proposed for on...
详细信息
To facilitate the integration of learning resources categorized under different ontology representations, the techniques of ontology mapping can be applied. Though many algorithms and systems have been proposed for ontology mapping, they do not have an automatic weighting strategy on class features to automate the ontology mapping process. A novel method of computing the feature weights is proposed. By feature semantic analysis, the different entities similarity calculation model and weight calculation model were defined. The results show that it makes the ontology mapping process more automatic while retaining satisfying accuracy. Improve ontology mapping effectiveness.
暂无评论