Detecting and locating surface defects in textured materials is a crucial but challenging problem due to factors such as texture variations and lack of adequate defective samples prior to testing. In this paper we present a novel unsupervised method for automatically detecting defects in fabrics based on a deep convolutional generative adversarial network (DCGAN). The proposed method extends the standard DCGAN, which consists of a discriminator and a generator, by introducing a new encoder component. With the assistance of this encoder, our model can reconstruct a given query image such that no defects but only normal textures will be preserved in the reconstruction. Therefore, when subtracting the reconstruction from the original image, a residual map can be created to highlight potential defective regions. Besides, our model generates a likelihood map for the image under inspection where each pixel value indicates the probability of occurrence of defects at that location. The residual map and the likelihood map are then synthesized together to form an enhanced fusion map. Typically, the fusion map exhibits uniform gray levels over defect-free regions but distinct deviations over defective areas, which can be further thresholded to produce a binarized segmentation result. Our model can be unsupervisedly trained by feeding with a set of small-sized image patches picked from a few defect-free examples. The training is divided into several successively performed stages, each under an individual training strategy. The performance of the proposed method has been extensively evaluated by a variety of real fabric samples. The experimental results in comparison with other methods demonstrate its effectiveness in fabric defect detection.
On the basis of the precise phase control of vanadium phosphorus oxides (VPOs), nanosized TiO 2 was employed as a dopant/dispersant to fabricate a series of VPO-TiO 2 catalysts through a wet mechanical co-milling process. The resulting catalysts showed outstanding durability plus excellent target products [acrylic acid (AA) + methyl acrylate (MA)] selectivity via acetic acid (HAc)−formaldehyde (FA) condensation. Over an optimized catalyst of 20% VPO-TiO 2 , the (AA + MA) selectivity being 85% (HAc input-based) at a yield level >60% (FA input-based) can be achieved after 180-h running, the best known to date over the VPObased catalysts. The detailed characterizations including X-ray powder diffraction, Raman spectra, XPS, and H 2 -TPR indicated that the V 5+ in the original VOPO 4 phase would be partially reduced in the presence of TiO 2 after the milling process in the cyclohexane medium; and the partially reduced VOPO 4 phase together with the decorated TiO 2 component stabilized the remaining V 5+ entities and considerably slowed down the continuous reduction of surface V 5+ species, accounting for substantially enhanced catalyst durability as well as target product selectivity. The NH 3 -/CO 2 -TPD results demonstrated that the surface acid−base property also varied notably with the VPO content which in turn controlled the HAc conversion and (MA + AA) selectivity accordingly.
A new type of supported vanadium phosphorus oxide (VPO) with self-phase regulation was simply fabricated (organic solvent free) for the first time by depositing the specific VPO precursor NH4(VO2)HPO4 onto the Siliceous Mesostructured Cellular Foams (MCF) with controlled activation. The resulting materials were found to be highly efficient and selective for sustainable acrylic acid (AA) plus methyl acrylate (MA) production via a condensation route between acetic acid (HAc) and formaldehyde (HCHO). A (AA + MA) yield of 83.7% (HCHO input-based) or a (AA + MA) selectivity of 81.7% (converted HAc-based) are achievable at 360 °C. The systematic characterizations and evaluations demonstrate a unique surface regulation occurring between the MCF and the NH4(VO2)HPO4 precursor. NH3 release upon activation of NH4(VO2)HPO4 precursor together with adsorption of NH3 by MCF automatically induces partial reduction of V5+ whose content is fine-tunable by the VPO loading. Such a functionalization simultaneously modifies phase constitution and surface acidity/basicity of catalyst, hence readily controls catalytic performance.
Dry reforming of methane has been systematically investigated over a series of x‐Co@SiO2‐y catalysts where x is the Co particle size ranging from 11.1 to 121.3 nm while y denotes the silica shell thickness ranging from 6.0 to 21.9 nm. Various techniques including TEM, XRD, H2‐TPR/‐TPD, XPS, BET, O2‐TPO, TG, and H2‐TPSR‐MS were employed to characterize physicochemical properties of catalysts. H2‐TPR and XPS results indicate that the core–shell interaction is dependent on the core size: the smaller the Co particle size is; the stronger the core–shell interaction. The investigations employing H2‐TRSR‐MS and XPS on the spent catalysts demonstrated that a fraction of metallic Co was re‐oxidized on a large‐core catalyst such as 121.3‐Co@SiO2‐72.2 during the reaction, and such oxidation leads to lower catalytic activity and stability. O2‐TPO results indicated that the catalyst with smaller core size caused significant coking. TG analysis together with TEM investigation on the used samples suggested that carbon deposition is notably core‐size‐dependent and responsible for deactivation of the small‐core catalyst. Among various core–shell structured catalysts, 27.8‐Co@SiO2‐14.3 showed superior activity and durability, owing to the well‐balanced property between coking and anti‐oxidation of Co cores.
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.