Heat-treated hardened tool steels are significantly utilised in die and mould manufacturing, automotive, aerospace, and precision engineering industries due to their excellent hardness, wear resistance, and dimensional stability. However, machining these materials remains a significant manufacturing challenge because of high cutting forces, elevated cutting temperatures, rapid tool wear, and deterioration of surface integrity. To address these challenges, this study suggests an integrated statistical and multi-criteria decision-making (MCDM) method for the multi-response optimisation of CNC turning parameters of heat-treated MDC-K tool steel ( 58 HRC). The cutting speed (175–275 m/min), feed rate (0.1–0.3 mm/rev), and approach angle (70–90°) were chosen as control factors, whereas cutting force (N), surface roughness (Ra), and tool–chip contact length (L) was selected as response variables. The Box–Behnken design of the response surface methodology (RSM) was used for modelling the process, and the analysis of variance (ANOVA) was used to assess the statistical significance of the factors. To solve the multi-response optimisation problem, an integrated Multiple Criteria Ranking by Alternative Trace (MCRAT) and Ranking the Alternatives by Perimeter Similarity (RAPS) approach was employed; this allowed the objective weighting of the outputs and a powerful ranking of the alternatives of the machining processes. It can be concluded from the results that the surface roughness and cutting force are most sensitive to feed rate, and cutting speed shows significant control over other characteristics of the interaction between the tool and chip. The best machining conditions were determined to be 225 m/min, 0.1 mm/rev and 80° cutting angles, which give minimum cutting force and better surface quality. The validation of the proposed framework using the agreement between the RSM-based desirability optimisation and the MCDM ranking showed the reliability of the proposed framework. This is an integrated approach that presents a systematic and efficient approach for optimising machining parameters for hard-to-machine materials, which may be implemented in high-tech manufacturing processes.

Machining of hardened tool steels concerns severe unstable surface integrity, tool wear, and complex thermomechanical interfaces, presenting challenges for sustainable manufacturing. This work aims to investigate the effect of nano-aluminium oxide (Al2O3) minimum quantity lubrication (MQL) on the machinability of heat-treated high-speed tool steel (YXR-7) and to evaluate the ability of data-driven models in predicting machining responses under limited experimental conditions. A total of 81 controlled milling experiments were performed to estimate surface roughness (Ra), material removal rate (MRR), and tool wear rate (TWR) under dry and nano-MQL conditions. Experimental findings confirmed that nano-MQL decreased surface roughness by 65.61% and tool wear rate by nearly 56% compared to dry machining, while maintaining higher productivity. FESEM (Field emission scanning electron microscopy) observations showed an evolution from severe adhesive-diffusive wear in dry cutting to moderately mild abrasive-oxidative wear with tribofilm formation under nano-lubrication. XRD (X-ray diffractometry) analysis confirmed the presence of oxide and various carbide phases, supporting the experience of stress-assisted tribo-chemical relations. To investigate predictive performance, Multiple Linear Regression (MLR), Extreme Gradient Boosting (XGBoost), Feedforward Neural Network (FNN), and Gaussian Process Regression (GPR) models were utilised by cross-validation. For the available dataset, simpler regression models exhibited stable generalisation behaviour, while more complex nonlinear models established sensitivity to data scale. The integrated experimental and modelling approach provides insight into lubrication-driven wear mechanisms and offers practical guidance for improving process efficiency and sustainability in hard milling applications.

Additive manufacturing through Laser Powder Bed Fusion (L-PBF) offers significant advantages for tooling applications. However, its implementation in high-load tribological conditions is limited by process-induced defects. In this study, the dry sliding wear behaviour of L-PBF fabricated maraging steel was scientifically investigated under high-load conditions (150 N), demonstrating industrial stamping operations and compared with conventionally manufactured AISI D2 tool steel. Both materials were heat-treated to attain a comparable hardness of 9.74 ± 0.43 GPa (~ 58–60 HRC) to enable a reasonable performance evaluation. The results reveal that the L-PBF maraging steel shows a significantly higher specific wear rate (on order of 10− 6 mm3/Nm) and a comparatively higher and less stable coefficient of friction than AISI D2 steel. Comprehensive FESEM analysis of worn surfaces shows that defects such as microporosity, lack of fusion, and microcracks act as stress concentrators and enable abrasive particle entrapment, leading to severe three-body abrasive wear. Nanoindentation results further confirm defect-induced deformation behaviour through distinct pop-in events, establishing a direct correlation between microstructural heterogeneity and tribological degradation. Similar to conventional low-load studies, the present high-load study highlights the critical role of defect-driven mechanisms in governing wear performance. The findings reveal that attaining high hardness alone is insufficient for ensuring wear resistance in L-PBF tool steels and highlight the necessity of defect control for reliable application in stamping die conditions.

Deepfake technology is generally used by malicious actors to create deceptive multimedia content. Unfortunately, these artificially generated contents pose significant harm and risk to society. The study organizes and categorizes various approaches, including techniques for image, video, audio and multimodal deepfake detection. This work also introduces a comparative study of state-of-the-art techniques and highlights the performance evaluation of existing deepfake detection techniques. Compared to existing deepfake surveys, this study focuses on emerging trends such as multimodal deepfake detection and real-time deepfake forensics. The insights of this work can potentially help identify gaps and refine existing approaches to develop more accurate and efficient deepfake detection solutions.

This study presents an integrated optimisation method for CNC milling of heat-treated YXR7 tool steel using carbide cutting inserts under varying lubrication and process parameters. A full factorial experimental design comprising 27 runs was employed to assess the influence of depth of cut (dc), feed per tooth (ft), cutting speed (Cs), and nano-cutting fluid (Cf) on critical performance responses such as surface roughness (Ra), material removal rate (MRR), and tool wear rate (TWR). An advanced modelling through regression and ANOVA showed complex interactive and non-linear effects among process parameters. To effectively navigate these interdependencies, a novel hybrid decision-making model combining the Full Consistency Method (FUCOM) and fuzzy-MARCOS was employed. This multi-criteria decision-making (MCDM) method was described for uncertainties in machining performance and successfully ranked experimental alternatives based on their proximity to ideal performance. The optimal configuration (Experiment 21) accomplished a superior balance across all criteria, notably achieving a low surface roughness (Ra ≈ 0.42 µm) and TWR (~0.148 mm³/min) while maintaining a high MRR (~109.4 mm³/min). The proposed fuzzy-FUCOM-MARCOS method reveals high robustness, adaptability, and decision reliability, contributing a valuable strategy for precision machining of hard-to-cut steels. This work bridges experimental understandings with intelligent optimisation, fostering sustainable and high-performance manufacturing practices in the tooling industry.

Aluminum and its alloys, such as Al–15Fe, Al–10Fe, and Al–5Fe, possess a unique blend of mechanical properties attributed to the development of intermetallic phases. These alloys are formed through the metallurgical casting process, where aluminum reacts chemically with iron. Renowned for their outstanding mechanical and physical characteristics, including corrosion resistance, high strength, thermal stability, and low density, Al–Fe intermetallic alloys find extensive applications across industries such as automotive, aerospace, chemical, medical, and electronics. This study involves the synthesis of Al–Fe intermetallic alloys through metallurgical casting using an induction furnace, supported by molecular dynamics simulations. The mechanical properties, including yield strength and hardness, were systematically assessed with varying iron content in the alloys. Results indicate a significant enhancement in these properties with increasing Fe content. Furthermore, structural phase analysis and the formation of intermetallic compounds were examined using optical microscopy and atomic-level investigations, offering a detailed understanding of the microstructural and intermetallic phase distribution within the materials.

H13 tool steel was fabricated by Laser Powder Bed Fused (LPBF) and the influence of annealing and tempering on tribological performance and microstructure was compared. The experiments produced as-built samples with heterogeneous melt pools, residual stresses, and a brittle martensitic matrix, which resulted in an unstable friction coefficient (0.63±0.01) and delamination wear. The XRD (X-ray diffraction) peak broadening and peak shifts showed lattice distortion and microstrain in the as-built condition, and they decreased after heat treatment, along with retained austenite reduction and carbide precipitation. Stable tribolayer formation resulted in the lowest friction coefficient (0.42±0.02) and smoother wear tracks in annealed samples, whereas the tempered samples had higher hardness (55±0.1 HRC) and two-body abrasion (0.58±0.02). The wear rates were 2.1✕10-4 mm3/Nm (as-built), 3.47✕10-3 mm3/Nm (annealed), and 1.6✕10-4 mm3/Nm (tempered). The results show that the correct heat treatment maximises the microstructure-tribology associations in LPBF H13, which enables its reuse in high-performance tooling.

Dies made of D2 tool steel commonly fail in heavy-duty cutting due to excessive wear from contact with sheet metal. Their wear resistance depends on steel composition, heat and surface treatment, and machining quality. Surface roughness from machining is a key factor in die failure. This study examines the impact of surfaces produced by various machining processes on the tribological properties of tool steel under lubricated and dry sliding conditions. Heat-treated D2 tool steel was machined using four processes-wire-cut electric discharge machining (WEDM), electric discharge machining (EDM), computer numerical control milling (CNC), and surface grinding (SG)-to generate surfaces with varying roughness. Surface roughness measurements revealed that CNC-milled surfaces had the highest roughness parameters, followed by SG, EDM, and WEDM. Based on the tribological results, artificial neural models (ANN) were used to predict the effect of roughness parameters on the coefficient of friction (COF) and specific wear rates (SWR). The ANN predictions were further validated using the novel MCDM approach. These approaches concluded that Ra, Rq, and Rv significantly influenced COF and SWR under lubricated conditions, whereas Ra, Rp, and Rv influenced the same under dry sliding conditions. Three-body abrasive wear was the dominant wear mechanism under both conditions. Additionally, micro-grooves were generated due to this wear mechanism under dry sliding conditions compared to lubricated ones.

The widespread circulation of manipulated videos using deepfake techniques has raised concerns about the authenticity of multimedia content. In response, deepfake detection techniques have made significant strides in specific scenarios. However, most of the existing methods are unimodal and focus only on extracting traditional spatial features, due to which they struggle to accurately identify modern deepfakes. This work introduces the STKD-VViT model for detecting deepfakes across multiple modalities while employing spatiotemporal features. STKD-VViT combines the strengths of the Video vision transformer and the Vision transformer to process visual and audio streams. The Video vision transformer employs a multi-head attention mechanism and tubelet embedding to extract the video’s spatial and temporal features. Alternatively, the vision transformer extracts the salient features from the mel-spectrograms of audio files. Furthermore, STKD-VViT leverages the knowledge distillation technique to reduce the number of FLOPs and the model’s parameters. Experimental results on the benchmark FakeAVCeleb dataset demonstrate that STKD-VViT achieves a testing accuracy of 97.49% for video stream data, 98.65% for audio stream data and 96.0% when both streams are combined using score-level fusion, surpassing other state-of-the-art methods.

The worldwide COVID-19 epidemic has emerged as a significant concern, affecting daily lives and underscoring the importance of early diagnosis for effective treatment in medical and healthcare settings. Current diagnostic testing for COVID-19 is sluggish, typically requiring hours to get results. Detection of COVID-19 from medical imaging presents a challenging task that has gained substantial interest from experts worldwide. Essential imaging modalities for diagnosing COVID-19 include chest X-rays and computed tomography (CT) scans. By contrast, most of the chest radiography can be completed in within fifteen minutes. Thus, employing chest radiography gives a possibility for early and reliable diagnosis of COVID-19, intending to relieve therapeutic obstacles for patients and speed up the diagnostic process. Recently, deep learning (DL) techniques have been shown to be effective in image-based diagnostics. This paper proposed an advanced deep convolution neural network (ADConv-Net) for COVID-19 detection and categorization using chest X-ray and CT images. The proposed technique is not only capable of recognizing critical connections and similarities in image classification, but also leads to improved diagnostic accuracy. The proposed model undergoes thorough evaluation for standard performance metrics. After evaluation, the ADConv-Net model achieves high accuracies of 98.84

Due to advancements in the deep learning technology, object detection has become significantly important for lane detection and vehicle detection. In recent times, lane detection has become more popular as it plays a significant role in traffic surveillance compared to other object detection technology. However, these strategies have several intrinsic flaws which need to be addressed. Traditional-based techniques still suffer from the challenges of the effectiveness and accuracy, whereas a complex convolutional layer is a challenge for deep learning-based strategies. A parameter selection issue affects the majority of the available lane detection algorithms, which further contributes to their unsatisfactory detection performance. In this study, we provide an effective lane detection method based on semantic segmentation to identify lane lines in a high-dimensional dataset by adding vertical spatial properties and contextual driving information. This paper employs two created frames—feature merging block and information exchange block—to identify unclear and obstructed lane lines more effectively. The simulations have been carried out for the proposed model on TUSimple and CULane datasets which resulted with 94.42

A commercially available DAC-10 tool steel substrates were plasma nitrided followed by vacuum heat treatment. Titanium nitride (TiN) with and without chromium (Cr) interlayer was deposited using a cathodic-arc deposition technique. The surface morphology of thin films was analyzed using X-ray diffractometry (XRD) and field emission scanning electron microscopy (FE-SEM) respectively. On both the TiN and Cr/TiN films, subsequent investigations comprising nanoindentation and nanoscratch tests were carried out. A notable 21.94 % increase in hardness was seen in the Cr/TiN film compared to TiN, plasma nitride, and heat-treated DAC-10 tool steel. There is an important variable that contributed to this improvement i.e., chromium (Cr) atoms had diffused into the TiN film from the Cr interlayer, which was also detected in the XRD pattern with (220) plane orientation. Along with an increase in hardness, the Cr/TiN film showed a 12.3 % increase in elastic modulus over the TiN film. Additionally, it was discovered that the strain hardening exponent was higher for the Cr/TiN film (0.38 vs. 0.33 TiN), indicating less pile-up formation during indentation in the Cr/TiN film. The Cr/TiN film’s scratch width and depth were reduced due to the higher scratch hardness. The projected wear rate of the scratched Cr/TiN film was significantly reduced by 71.23 % as a result of the coefficient of friction decreasing from 0.45 ± 0.05 for the TiN film to 0.33 ± 0.04 for the Cr/TiN film. These results show surface modification was the best alternative to enhance the nanomechanical and tribological properties of DAC-10 tool steel.

Usually heat treated tool steel is further processed with hard ceramic-based coatings to improve the wear resistance and mechanical properties of the die elements. Over the years, a multitude of coating materials have been developed, however, the right selection of the coating material remains a challenge. The present work examines the coating material selection problem for sheet metal forming applications using mathematical models from four different multi-criteria decision-making (MCDM) approaches merged with the MEREC (Method Based on the Removal Effects of Criteria) method. Criterion weights are determined by the MEREC technique, and nine different types of coating material choices are ranked by the MCDM methods. The robustness of the rankings, was further confirmed using sensitivity analysis carried out in four steps. The resilience and dependability of the TOPSIS approach in resolving MCDM issues are demonstrated by the sensitivity analysis results. The study also employs the suggested methodology for a coating material selection problem that has been published previously, producing rankings that are generally consistent with those reported in the literature.

Investigating the quantitative effects of feed rate (Fr), cutting speed (Cv), depth of cut (Dc) and nose radius (Nr) on the surface quality of the heat-treated tool steel workpiece and the tool wear of the Cr-(CrN/TiN) coated insert required the use of a response surface methodology (RSM). Initially, the experimental work has been conducted based on the L27 design of the experiment. Based on the experimental results, RSM has been applied to study the effect of turning process parameters on the responses namely the surface roughness of turned tool steel (Ra), and tool wear rate (TWR). To comprehend the interactive impact of the process variables on the quality indicators, ANOVA analysis and regression were used. Moreover, the machining process parameters were optimized using a novel MEREC-integrated fuzzy MARCOS approach, where the MEREC method was used to determine the criteria weight while the fuzzy MARCOS approach determined the ranking of the parameters. Based on the ranking results suitable set of machining process parameters was selected, i.e., (Fr = 0.3 mm/rev, Cv = 145 mm/min, Dc = 0.4 mm, Nr = 0.6 mm) which gives desirable outputs (TWR—0.1301 mm3/min and Ra—0.4617 µm). The robustness of the ranking obtained from the above-mentioned methodology was tested using different sensitivity analyses. Finally, a confirmation test was done for the optimized parametric set of experiments to validate the result.

Accurate segmentation of brain tumor regions in MRI images is essential for monitoring tumor growth. In view of this, several automated brain tumor segmentation models are proposed. U-Net is one of the most popular models widely used for image segmentation. Based on the U-Net architecture, complex models are proposed for brain tumor segmentation. However, existing models mainly focus on increasing the depth of the model for better accuracy without much bothering the transmission of spatial information. Attention-based U-Net model can help to extract salient local-level features that can be passed to the decoder part of the network. Considering the above issues, we propose a variant of the U-Net model called the Attention-based Residual Light U-Net model. The proposed model is effective in the sense that it has comparatively less model complexity, a better attention mechanism, and a combined loss function that can handle class imbalance problems. Our experiments on one of the latest BraTS-2021 datasets show that the proposed method can achieve a mean Intersection over Union of 0.8874, 0.8863, 0.8905, and dice score of 0.9415, 0.9387, 0.9115 on the whole tumor, tumor core, and enhancing tumor, respectively.