Additive manufacturing (AM), commonly known as 3D printing, is a rapidly growing technology. Guaranteeing the quality and mechanical strength of printed parts an active research area. Most existing methods adopt open-loop-like Machine Learning (ML) algorithms that can be used only for predicting properties without any assuring mechanism. Some closed-loop approaches, on other hand, consider single adjustable processing parameter to monitor part. This work proposes both open-loop ML models...

Additive manufacturing (AM) can build up complex parts in a layer-by-layer manner, which is kind of novel and flexible production technology. The special capability AM shows great application potential various fields. However, an open-loop control method cannot guarantee the reliability repeatability process. Defects often occur to deteriorate product quality lead material time waste, hinders development industry. In this regard, lot efforts have been made make process more controllable....

A radiance field is an effective representation of 3D scenes, which has been widely adopted in novel-view synthesis and reconstruction. It still open challenging problem to evaluate the geometry, i.e., density field, as ground-truth almost impossible obtain. One alternative indirect solution transform into a point-cloud compute its Chamfer Distance with scanned ground-truth. However, many widely-used datasets have no since scanning process along equipment expensive complicated. To this end,...

The additive manufacturing technologies develop very fast in recent years. However, we still face the problem of lacking online feedback for system, and reliability robustness need to be improved. In this paper, propose concept parallel systems, which combine with manufacturing. Furthermore, present how use artificial systems-computational experiments-parallel execution approach make 3D printers intelligent, order handle uncertainties unexpected conditions.

A radiance field is an effective representation of 3D scenes, which has been widely adopted in novel-view synthesis and reconstruction. It still open challenging problem to evaluate the geometry, i.e., density field, as ground-truth almost impossible obtain. One alternative indirect solution transform into a point-cloud compute its Chamfer Distance with scanned ground-truth. However, many widely-used datasets have no since scanning process along equipment expensive complicated. To this end,...

Fault diagnosis is a vital technique to pinpoint the machine malfunctions in manufacturing systems. In recent years, deep learning techniques greatly improve fault detection accuracy, but there still remain some problems. If one absent training data or signal disturbed by severe noise interference, classifier may misjudge health state. This problem limits reliability of real applications. this paper, we enhance method using Bayesian Convolutional Neural Network (BCNN). A Shannon...

Abstract Additive manufacturing (AM) is gaining prominence across numerous fields, which involves the generation of extensive data at each process stage. A relational database a useful tool to store such AM and streamline retrieval. Users can specify value one variable or attribute retrieve corresponding record values another attribute. This establishes correlations between variables, supports applications as planning. Nonetheless, an operation “hard” query, lacks reasoning capabilities...

Novel views synthesis is an important topic in metaverse application. Existing methods suffer the tremendous training to guarantee quality, which a stringent condition practice. To address this problem, we propose depth-guided and self-supervised method achieving novel challenging sparse views. For goal, depth information digging strategy uncertainty-based supervision method. We conduct series of experiments on DTU dataset demonstrate rationality our design. And experiment results represent...

Due to the scarcity of point cloud datasets in a specific domain, utilizing generative model approaches becomes essential for data augmentation. Diffusion models have demonstrated impressive capabilities generation through guided reverse process. In this work, we employ process Markov chain conditioned on shape latent progressively generate dental crown from noise distribution. We propose map global set partlevel implicit representations and introduce cross-attention block provide geometric...

Audio-driven talking head generation is a pivotal area within film-making and Virtual Reality. Although existing methods have made significant strides following the end-to-end paradigm, they still encounter challenges in producing videos with high-frequency details due to their limited expressivity this domain. This limitation has prompted us explore an effective post-processing approach synthesize photo-realistic videos. Specifically, we employ pretrained Wav2Lip model as our foundation...

Recent advancements in models linking natural language with human motions have shown significant promise motion generation and editing based on instructional text. Motivated by applications sports coaching motor skill learning, we investigate the inverse problem: generating corrective text, leveraging models. We introduce a novel approach that, given user’s current (source) desired (target), generates text instructions to guide user towards achieving target motion. leverage large generate...

Recent advancements in models linking natural language with human motions have shown significant promise motion generation and editing based on instructional text. Motivated by applications sports coaching motor skill learning, we investigate the inverse problem: generating corrective text, leveraging models. We introduce a novel approach that, given user's current (source) desired (target), generates text instructions to guide user towards achieving target motion. leverage large generate...