Additive manufacturing technologies can produce complex components in a more efficient and environmentally sustainable way when compared with conventional manufacturing technologies. These advantages are expected to keep cementing Additive manufacturing role in several industries. Additive manufacturing technologies are fast-improving technologies with an estimated doubling of performance on average every five years. These fast developments make it difficult to keep track of current technological capabilities and environmental impacts. Among the most relevant additive manufacturing technologies there are the metal additive technologies of Powder Bed Fusion, Directed Energy Deposition and Binder Jetting. This paper assesses the current technological capabilities and environmental impact of these technologies. For that, resorting to data from literature, technical reports and company data sheets, the technological steps are defined, the relevant process parameters are established, and qualitative and quantitative data is gathered. A set of visualizations of process capabilities and their environmental impact is performed with the intent of helping designers and engineers with decision-making. It is also interesting for the research community to visualize and understand the current capabilities and to establish roadmaps for research. Limitations of this study include data currently available for each parameter, and the shortage of data for the environmental calculations.

This study empirically estimates the technological improvement rate (TIR) of metal powder bed fusion (PBF) technology, widely used in aerospace, automotive, and medical industries. PBF's continuous long-term adoption growth is driven by its ability to enhance manufacturing efficiency in terms of time and raw material use, as well as its capability to produce high-quality, high-strength, complex-shaped parts. Measuring the technological development of PBF is crucial as itis enlarging its application domain and is increasingly considered a viable alternative to traditional manufacturing technologies across a broader range of applications. We resorted to the literature to collect information and assess which technical parameters are most relevant to measure the capabilities of PBF. With those, we established an ideal functional performance metric (FPM) capable of comprehensively assessing PBF's technological performance improvement. Considering all available data sources and PBF machines ever made commercially available, a data set of technical parameters was constructed. This was followed by a data curation process focusing on data availability and reliability. The resultant practical FPM was used to estimate the TIR of PBF technology. By employing regression analysis, we estimate a yearly improvement of 26.8%. This empirical rate comes as a more accurate and reliable substitute to the previously indirectly estimated patent-derived rate of 33.3%. Our findings underscore PBF's capability of keeping pace with its growing significance and wider industrial applications. The results of this study provide a key metric for those in the industry and research, confirming the rapid performance growth and establishing a standard for future industrial uses.

Deploying conditional Generative Adversarial Networks (cGANs) for industrial texture synthesis faces two barriers: the prohibitive cost of manual data annotation and the uncertain alignment between automated evaluation metrics and human perception. This study addresses both challenges for marble texture synthesis using 289 high-resolution industrial scans. We adapt an unsupervised segmentation pipeline combining Simple Linear Iterative Clustering (SLIC) superpixels, Gaussian Mixture Models (GMMs), and graph cut optimization to extract vein structures without manual annotation. Four cGAN architectures—baseline cGAN, Pix2Pix, BicycleGAN, and GauGAN—are benchmarked using a dual-evaluation protocol contrasting ten automated metrics with structured human-centered assessment. The results reveal a significant metric–perception discrepancy. Pix2Pix achieved the best Fréchet Inception Distance (FID = 85.3) yet received the lowest human ratings due to periodic texture artifacts. GauGAN produced textures statistically indistinguishable from real marble, achieving a Visual Turing Pass Rate (VTPR) of 0.533 and a Mean Opinion Score on Marble Authenticity (MOS-MA) of 2.89, despite an inferior FID (87.3). These findings make three contributions: an annotation-free segmentation pipeline, empirical evidence that automated metrics alone are insufficient for architecture selection, and a dual-evaluation framework that establishes human-in-the-loop assessment as essential for quality-critical industrial deployment.

The marble industry relies on proprietary commercial names rather than objective visual categories, creating market inefficiencies for stakeholders who select stones based on appearance. Supervised classification methods perpetuate this problem by replicating inconsistent commercial labels instead of discovering intrinsic visual structure. We propose an unsupervised pipeline combining a two-stage training strategy, pure self-supervised pretraining followed by cluster-aware fine-tuning of a DINO Vision Transformer, with UMAP dimensionality reduction and Ward's agglomerative hierarchical clustering. Systematic ablation studies on 1,540 marble images spanning 10 commercial varieties validate each design choice: cluster-aware training at k=10 yields superior embeddings over the self-supervised baseline (Silhouette Score 0.778 vs. 0.761; Davies–Bouldin Index 0.293 vs. 0.364), UMAP compression to five dimensions resolves high-dimensional noise pathologies, and Ward's linkage produces the most compact partitions. The resulting taxonomy reveals three phenomena invisible to commercial classification: cross-category merging of visually indistinguishable stones carrying different market names, intra-category splitting of heterogeneous sub-populations within single varieties, and coherent grouping where commercial and visual boundaries coincide. We further demonstrate that standard extrinsic metrics are misaligned with unsupervised taxonomy objectives when reference labels encode the inconsistencies the method aims to resolve. This work provides a validated methodology for data-driven visual classification in the natural stone industry and a transferable template for domains with unreliable labelling conventions.