Quad meshes are essential in geometric modeling and computational mechanics. Although learning-based methods for triangle mesh demonstrate considerable advancements, quad mesh generation remains less explored due to the challenge of ensuring coplanarity, convexity, and quad-only meshes. In this paper, we presentPoint2Quad, the first learning-based method for quad-only mesh generation from point clouds. The key idea is learning to identify quad mesh with fused pointwise and facewise features. Specifically, Point2Quad begins with a k-NN-based candidate generation considering the coplanarity and squareness. Then, two encoders are followed to extract geometric and topological features that address the challenge of quad-related constraints, especially by combining in-depth quadrilaterals-specific characteristics. Subsequently, the extracted features are fused to train the classifier with a designed compound loss. The final results are derived after the refinement by a quad-specific post-processing. Extensive experiments on both clear and noise data demonstrate the effectiveness and superiority of Point2Quad, compared to baseline methods under comprehensive metrics. The code and dataset are available athttps://github.com/cognaclee/Point2Quad.

Abstract By transmitting 2‐bits orthogonal information signal in parallel, compared with phase‐orthogonality CDSK (PO‐CDSK), high‐data‐rate PO‐CDSK (HPO‐CDSK) has double spectral efficiency. Unfortunately, the additional information will cause more noise interference during correlation detection in HPO‐CDSK. The study of HPO‐CDSK is extended and improved HPO‐CDSK (I‐HPO‐CDSK) is proposed to obtain high reliability and effectiveness. By virtue of the theorem that the integral of the sine and cosine functions in a period is 0, in I‐HPO‐CDSK, 3‐bits information can be transmitted in parallel and will not cause ISI components. Moreover, compared with HPO‐CDSK, I‐HPO‐CDSK can decrease noise interference. Results show that, due to less interference and 3‐bits information transmitting in parallel, I‐HPO‐CDSK has better BER performance and spectral efficiency than those of previous chaotic communication schemes, such as PO‐CDSK and HPO‐CDSK.

To eliminate the intrasignal interference (ISI), phase‐orthogonality correlation‐delay‐shift‐keying (PO‐CDSK) alternately chooses quadrature sinusoidal wavelets to modulate the reference signals in different symbol durations and keeps the reference and information‐bearing signals orthogonal in one symbol duration. Unfortunately, in PO‐CDSK, one data‐modulated signal can carry only 1‐bit information, resulting in a relatively low data rate. In this brief, the authors design a high‐data‐rate PO‐CDSK system, in which 2‐bits information share one time slot but transmitted over the inphase and quadrature channels, respectively. Theoretical BER expression of the proposed system is studied analytically, and simulations are performed over additive white Gaussian noise and Rayleigh multipath fading channels for verification. Results show that, without ISI components, the authors' new system performs much better than differential‐chaos‐shift‐keying (DCSK), CDSK and high‐efficiency DCSK systems in BER performance. Besides, the proposed system also achieves doubled spectral and energy efficiencies in comparison with PO‐CDSK.

To overcome the drawback of bit‐error rate (BER) performance decrement caused by intrasignal interference (ISI) in correlation‐delay‐shift‐keying (CDSK) and its enhanced schemes, in this brief, a reliable and effective chaotic communication scheme: phase‐orthogonality CDSK is proposed. In the proposed scheme, quadrature sinusoidal wavelets are alternately selected to modulate the reference signal. This operation makes the reference signal strictly orthogonal to the information‐bearing signal and the ISI component is eliminated during the demodulation. The BER performance is analytically studied and the simulations are performed in additive white Gaussian noise channel and Rayleigh multipath fading channels. Theoretical analysis and simulations show that, since there is no ISI component and better bandwidth efficiency, the proposed BER performance is much better than differential‐chaos‐shift‐keying (DCSK) and CDSK. With the length of multiframe increasing, the proposed BER performance gradually becomes better, and can even outperform reference‐adaptive CDSK and improved DCSK.