Search

The detection of structure centerlines from imaging data plays a crucial role in the understanding, application and further analysis of many diverse problems, such as road mapping, crack detection, medical imaging and biometric identification. In each of these cases, pixel-wise segmentation is not sufficient to understand and quantify overall graph structure and connectivity without further processing that can lead to compound error. We thus require a method for automatic extraction of graph representations of patterning. In this paper, we propose a novel Deep Orientated Distance-transform Network (DODN), which predicts the centerline map and an orientated distance map, comprising orientation and distance in relation to the centerline and allowing exploitation of its geometric properties. This is refined by jointly modeling the relationship between neighboring pixels and connectivity to further enhance the estimated centerline and produce a graph of the structure. The proposed approach is evaluated on a diverse range of problems, including crack detection, road mapping and superficial vein centerline detection from infrared/ color images, improving over the state-of-the-art by 2.1%, 10.9% and 17.3%/ 4.6% respectively in terms of quality, demonstrating its generalizability and performance in a wide range of mapping problems. • We present a novel multitask learning architecture for centerline detection. • We incorporate task uncertainty to simultaneously learn multiple losses. • A graph-cut based connectivity refinement approach is proposed. • Our proposed approach outperforms state-of-the-art methods on three datasets. [ABSTRACT FROM AUTHOR]

The implementation of extrusion-based 3D concrete printing (3DCP) in large-scale constructions is currently limited by concerns regarding rheology control and the sustainability of this process. To address these issues, this study presents an approach to develop limestone-calcined clay-based cementitious (LC3) materials accelerated by Ca(NO 3) 2 solution in an inline static mixer-based 3DCP setup. Using this approach, a printable mixture containing only about 275 kg/m3 of Portland cement was formulated that can exhibit a good buildability performance and a 28-day compressive strength of over 30 MPa. Additionally, the effects of adding Ca(NO 3) 2 solution on the initial setting time, structural build-up, inline buildability, early-age hydration, and compressive strength of LC3 materials were investigated and discussed. Results show that the addition of Ca(NO 3) 2 solution improved the buildability and accelerated initial setting as well as the structuration due to the promoted ettringite precipitation and C –S–H nucleation. Furthermore, compressive strength at 7 and 28 days was improved through increasing the Ca(NO 3) 2 dosage, which can be attributed to the formation of NO 3 -AFm and the increase in the amount of C –S–H gels. [ABSTRACT FROM AUTHOR]

We propose an image-based technique to measure the volume, weight and the size distribution of gravel particles in a gravel-soil mixture. The proposed method uses 3D scanning and a surface reconstruction algorithm to generate a high-resolution depth image, which is then used to accurately estimate the volume and weight of each gravel particle. The proposed method is evaluated on several gravel soil samples collected from 25 farming locations. The experimental results show that the proposed technique produces an accurate estimate of gravel volumes and gravel weights. It achieves a relative root mean square error of 4% for large gravel particles and an overall correlation of 0.99 with the ground truth, for the task of gravel volume estimation. For the estimation of gravel weight distribution, the proposed method can reach a low root mean square error of 0.54%. The rapid measurement of the full spectrum of coarse fragments in soil, using this method, is an advantage compared to the manual methods. • A method to measure size distribution of gravel particles in a gravel-soil mixture. • It is an automatic image-processing and machine learning based method. • It can accurately estimate the volume and weight of each gravel particle. • The experimental results show the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Point-of-interest (POI) recommendation is an essential service to location-based social networks (LBSNs), benefiting both users providing them the chance to explore new locations and businesses by discovering new potential customers. These systems learn the preferences of users and their mobility patterns to generate relevant POI recommendations. Previous studies have shown that incorporating contextual information such as geographical, temporal, social, and categorical substantially improves the quality of POI recommendations. However, fewer works have studied in-depth the multi-aspect benefits of context fusion on POI recommendation, in particular on beyond-accuracy , fairness , and interpretability of recommendations. In this work, we propose a linear regression-based fusion of POI contexts that effectively finds the best combination of contexts for each (i) user, or (ii) group of users from their historical interactions. The results of large-scale experiments on two popular datasets Gowalla and Yelp reveal several interesting findings. First, the proposed approach does not present significant loss in accuracy and unfairness of popularity bias as with classical collaborative baselines, and yet improves the beyond-accuracy of recommendation compared with existing context-aware (CA) approaches using heuristic context fusions; for instance, the proposed approach improves the accuracy and beyond-accuracy compare to best baseline model by 25% and 30%, respectively. Second, our proposed approach is interpretable, allowing to explain to the user why she has been recommended specific POIs, based on the learned context weights from user past check-ins; for example, if you are in Rome and our method recommends you a historical place like 'Colosseum', it can also provide an explanation why this item is recommended to you based on your personal preference on context (e.g., you were recommended to visit 'Colosseum' because in the past your visited historical places). Third, by analyzing the fairness of recommendation with respect to users (based on their activity levels) and items (based on the popularity of items), we found that a model which is recommend fairly on one dataset can recommend unfair on another dataset. Overall, our study suggests that appropriate context fusion is an essential element of an accurate, fair, and transparent POI recommendation system. We highlight that while we have tested the efficacy of our context-fusion methods on two popular CA recommendation models in the POI domain, namely GeoSoCa and LORE, our system can be flexibly utilized to extend the capability of other CA algorithms. • The paper investigates the role of context fusion in POI recommendations. • Evaluations of the accuracy, beyond-accuracy, and fairness of POI recommendations. • We provide better recommendations in terms of beyond accuracy and accuracy metrics. • The study pushes the research forward by choosing several evaluation objectives. • The results of fairness show that contextual models provide better fairness results. [ABSTRACT FROM AUTHOR]

We propose a Locality-constrained Linear Coding (LLC) based algorithm that captures discriminative information of human actions in spatio-temporal subsequences of videos. The input video is divided into equally spaced overlapping spatio-temporal subsequences. Each subsequence is further divided into blocks and then cells. The spatio-temporal information in each cell is represented by a Histogram of Oriented 3D Gradients (HOG3D). LLC is then used to encode each block. We show that LLC gives more stable and repetitive codes compared to the standard Sparse Coding. The final representation of a video sequence is obtained using logistic regression with ℓ 2 regularization and classification is performed by a linear SVM. The proposed algorithm is applicable to conventional and depth videos. Experimental comparison with ten state-of-the-art methods on three depth video and two conventional video databases shows that the proposed method consistently achieves the best performance. [ABSTRACT FROM AUTHOR]

In this study, a solution for anisotropic conductive heat transfer in a composite cylindrical shell is computed by an exact analytical method. The fibers are considered to be winded around the cylinder at any arbitrary angle. A general linear boundary condition is applied on both circular bases of the cylindrical shell to account for various combinations of thermal conditions. The solution considers the effects of convection of the ambient flow motion and various external radiative heat fluxes. The analytical solution describes the temperature distribution in the axial and circumferential directions. In principle, the heat conduction equation should involve a dual second-order derivation, which precludes solving the equation by the direct application of common exact methods. Therefore, an appropriate canonical mapping is selected as a solution to cancel the dual derivation of temperature in the mapped equations. The separation of variables method is then applied to the mapped equations, which are conducted to obtain an exact form of the solution. Finally, an analytical technique based on the Fourier series concept is proposed to determine unknown coefficients in the final solution. The obtained analytical results correspond well with the solved numerical results. The capability of the current solution is examined by application to a few relevant industrial cases. [ABSTRACT FROM AUTHOR]

Predicting novel drug side-effects, or Adverse Drug Reactions (ADRs), plays an important role in the drug discovery process. Existing methods consider mainly the chemical and biological characteristics of each drug individually, thereby neglecting information hidden in the relationships among drugs. Complementary to the existing individual methods, in this paper, we propose a novel network approach for ADR prediction that is called Augmented Random-WAlk with Restarts (ARWAR). ARWAR, first, applies an existing method to build a network of highly related drugs. Then, it augments the original drug network by adding new nodes and new edges to the network and finally, it applies Random Walks with Restarts to predict novel ADRs. Empirical results show that the ARWAR method presented here outperforms the existing network approach by 20% with respect to average Fmeasure. Furthermore, ARWAR is capable of generating novel hypotheses about drugs with respect to novel and biologically meaningful ADR. [ABSTRACT FROM AUTHOR]

Graphene oxide (GO) was functionalized using three different diamines, namely ethylenediamine (EDA), 4,4′-diaminodiphenyl sulfone (DDS) and p -phenylenediamine (PPD) to reinforce an epoxy adhesive, with the aim of improving the bonding strength of carbon fiber/epoxy composite. The chemical structure of the functionalized GO (FGO) nanosheets was characterized by elemental analysis, FT-IR and XRD. Hand lay-up, as a simple method, was applied for 3-ply composite fabrication. In the sample preparation, the fiber-to-resin ratio of 40:60 (w:w) and fiber orientations of 0°, 90°, and 0° were used. The GO and FGO nanoparticles were first dispersed in the epoxy resin, and then the GO and FGO reinforced epoxy (GO- or FGO-epoxy) were directly introduced into the carbon fiber layers to improve the mechanical properties. The GO and FGO contents varied in the range of 0.1–0.5 wt%. Results showed that the mechanical properties, in terms of tensile and flexural properties, were mainly dependent on the type of GO functionalization followed by the percentage of modified GO. As a result, both the tensile and flexural strengths are effectively enhanced by the FGOs addition. The tensile and flexural moduli are also increased by the FGO filling in the epoxy resin due to the excellent elastic modulus of FGO. The optimal FGO content for effectively improving the overall composite mechanical performance was found to be 0.3 wt%. Scanning electron microscopy (SEM) revealed that the failure mechanism of carbon fibers pulled out from the epoxy matrix contributed to the enhancement of the mechanical performance of the epoxy. These results show that diamine FGOs can strengthen the interfacial bonding between the carbon fibers and the epoxy adhesive. [ABSTRACT FROM AUTHOR]

Abstract: Disk scheduling is an operating system process to service disk requests. It has an important role in QOS guarantee of soft real-time environments such as video-on-demand and multimedia servers. Since now, some disk scheduling algorithms have been proposed to schedule disk requests in an optimized manner. Most of these methods try to minimize makespan by decreasing the number of disk head seeks as one of the slowest operations in modern computers and crucial for system performance because it usually takes some milli-seconds. In this paper, we propose a new disk scheduling method based on genetic algorithm that considers makespan and number of missed tasks simultaneously. In the proposed method, a new coding scheme is presented which employs simple GA procedures such as crossover and mutation and a penalty function in fitness. To get the best performance of the proposed method, its parameters such as number of chromosomes in initial population, mutation, and crossover probabilities, etc have been adjusted by applying it on some sample problems. The algorithm has been tested on several problems and its results were compared with well-known related methods. Experimental results showed that the proposed method worked very well and excelled most related works in terms of miss ratio and average seeks. [Copyright &y& Elsevier]

• Chitosan based materials (CBMs) have great potential application in wound healing. • CBMs have great biodegradable, biocompatible and antimicrobial activities. • Several approaches have been presented to modify chitosan using amino acids. • Amino acids provides more re-epithelization and cell adhesion deposition. Chitosan, a well-known biopolymer due to its unique properties, has received considerable attention as a result of the amine group activity that locates on the backbone of chitosan. To improve the mechanical and antibacterial characteristics of chitosan, various modifications have been used. Amino acids, the monomeric units of proteins, among all other compounds have been chosen to discuss as promising materials for wound healing in combination with chitosan. This review aims to provide an up-to-date overview of the methods used for modification of chitosan and the potential biomedical application, in particular wound healing, reported in the literature during the last five years. [ABSTRACT FROM AUTHOR]

Chaos is one of the most significant findings in physics and engineering. Fractional-order chaotic systems are nonlinear systems with special features which, some of them include sensitivity to initial condition and the order of fractional derivative, unpredictable and complex dynamic behavior, high bandwidth and controllable noise-like behavior. By accepting this truth that any irregular behavior in a dynamic system is a sign of chaos and that a chaotic system is a deterministic system with pseudo-random behavior, chaos can be observed in different fields of science and engineering like mathematics, physics, mechanical, chemical and electrical engineering. Recently, control of fractional-order chaotic systems has been one of the most interesting topics that attracted many researchers' idea. This paper deals with the comprehensive study of control and synchronization of fractional order chaotic systems, and shows how chaos is formed in developing inter-disciplinary researches of fractional order systems from the first research till today. [ABSTRACT FROM AUTHOR]

Regularization is an essential aspect in the context of deep learning as it mitigates the risk of overfitting in deep neural networks. This study presents a novel deep learning regularization method, referred to DL-Reg, which effectively reduces the nonlinearity of deep networks by enforcing linearity to a certain extent. The proposed method is based on the incorporation of a linear constraint into the objective function of deep neural networks, which is defined as the error of a linear mapping from the inputs to the outputs of the model. Specifically, DL-Reg imposes a linear constraint on the network, which is further adjusted by a regularization factor, thereby preventing the network from overfitting. The effectiveness of DL-Reg is evaluated by training state-of-the-art deep network models on several benchmark datasets. The results of the experiments demonstrate that the proposed regularization method provides significant improvements over existing regularization techniques and enhances the performance of deep neural networks, particularly when dealing with small-sized training datasets. The main code of DL-Reg written in PyTorch is also available here: https://github.com/m2dgithub/DL-Reg.git. • Regularizing deep networks using a linear regression model. • Allowing supervised deep networks to be trained using the semi-supervised learning. • Increasing the performance of deep networks on the small-sized dataset. • Conducting extensive experiments to certify the significance of the proposed method. [ABSTRACT FROM AUTHOR]

• Training of a deep neural network for dehazing with unmatched training data. • Improvements over existing methods from based on objective PSNR and SSIM measures. • Demonstration of training method that can be applied to vision tasks. • Demonstration of effects of Haze on Object Detection Models. The rapid growth in computer vision applications that are affected by environmental conditions challenge the limitations of existing techniques. This is driving the development of new deep learning based vision techniques that are robust to environmental noise and interference. We propose a novel deep CNN model, which is trained from unmatched images for the purpose of image dehazing. This solution is enabled by the concept of the Siamese network architecture. Using object performance measures of image PSNR and SSIM we are able to demonstrate a quantitative and qualitative improvement in the network dehazing performance. This superior performance is achieved with significantly smaller training datasets than existing methods. [ABSTRACT FROM AUTHOR]

Despite extensive research, 3D face reconstruction from a single image remains an open research problem due to the high degree of variability in pose, occlusions and complex lighting conditions. While deep learning-based methods have achieved great success, they are usually limited to near frontal images and images that are free of occlusions. Also, the lack of diverse training data with 3D annotations considerably limits the performance of such methods. As such, existing methods fail to recover, with high fidelity, the facial details especially when dealing with images captured under extreme conditions. To address this issue, we propose an unsupervised coarse-to-fine framework for the reconstruction of 3D faces with detailed textures. Our core idea is that multiple images of the same person but captured under different viewing conditions should provide the same 3D face. We thus propose to leverage a self-augmentation learning technique to train a model that is robust to diverse variations. In addition, instead of directly employing image pixels, we use a set of discriminative features describing the identity and attributes of the face as input to the refinement module, making the model invariant to viewing conditions. This combination of self-augmentation learning with rich face-related features allows the reconstruction of plausible facial details even under challenging viewing conditions. We train the model end-to-end and in a self-supervised manner, without any 3D annotations, landmarks or identity labels, using a combination of an image-level photometric loss and a perception-level loss that is identity and attribute-aware. We evaluate the proposed approach on CelebA and AFLW2000 datasets, and demonstrate its robustness to appearance variations despite learning from unlabeled images. The qualitative comparisons indicate that our method produces detailed 3D faces even under extreme occlusions, out of plane rotations and noise perturbations where existing state-of-the-art methods often fail. We also quantitatively show that our method outperforms SOTA with more than 30.14%, 9.87% and 11.3% in terms of PSNR, SSIM and IDentity similarity, respectively. [ABSTRACT FROM AUTHOR]

Graphene oxide (GO) was functionalized using three different diamines, namely ethylenediamine (EDA), 4,4′-diaminodiphenyl sulfone (DDS) and p -phenylenediamine (PPD) to reinforce an epoxy/glass fiber (EP/GF) composite laminate, with the aim of improving the overall composite mechanical performance. Different mechanical characterization techniques were used to determine the mechanical performance, including: tensile stress strain, double cantilever beam (DCB) mode-I fracture toughness and dynamic mechanical thermal analysis (DMTA). Scanning electron microscopy (SEM) was used to support the results and conclusions. The results demonstrated remarkable enhancements in the mechanical performance of EP/GF composite laminates by incorporation of functionalized graphene oxide (FGO) nanofiller, whilst the mechanical performance of the GO reinforced composite only improved marginally. Finally, the mechanical performance of the EP/GF/FGO multi-scale composites was found to be dependent on the type of FGO functional groups; of which EDA exhibited the highest performance. These investigations confirmed that the EDA-FGO-reinforced EP/GF composites possess excellent potential to be used as multifunctional engineering materials in industrial applications. [ABSTRACT FROM AUTHOR]

Abstract: This study was conducted to determine the replacement value of faba bean meal (FBM) for soybean meal (SBM) in practical diets of Oncorhynchus mykiss fingerlings. Four isonitrogenous (42% crude protein) and isolipidic (20% total lipid) diets were formulated using commercial ingredients. Faba bean meal was added in the diets at levels of 0% (control) and 15, 30 and 45%. When the FBM content was increased in the diets, the soybean meal, fish meal and wheat flour were gradually reduced to adjust protein and carbohydrate contents. A completely randomized experimental design was developed with four treatments and three replicates. Two hundred and forty rainbow trouts with average initial weight of 12.5±0.09g were assigned to twelve experimental tanks. The experiment lasted for seven weeks. Based on results, survival did not differ among treatments (p>0.05). Growth performance and feed conversion ratio of rainbow trout fingerlings fed diets containing 15 and 30% FBM were similar (p>0.05) to the control diet. However, 15% FBM diet showed better performance (p<0.05) compared to the 30% FBM diet. The poorest growth performance and FCR were observed in the fish fed the diet containing 45% FBM. The apparent digestibility of dry matter and protein decreased with increasing levels of FBM from 30% to 45% in diet. No significant difference (p>0.05) was found in carcass protein and ash contents among the dietary treatments. The present results demonstrated that rainbow trout fingerlings can tolerate up to 30% FBM in diet, but maximum growth performance was found at the 15% FBM diet. [Copyright &y& Elsevier]

• Heat conduction in heterogeneous composite conical shells is solved analytically. • Fiber angle varies from zero to 90 degrees. • Heat conduction coefficients are temperature-depended. • Heat convection and external radiation are modeled. In present paper, problem of anisotropic heat conduction in heterogeneous composite conical shells is solved analytically. Arbitrary values for fiber angle (ranging from zero to 90 degrees) cause anisotropy for the heat conduction problem in composite conical shells. In our analysis, heat convection between conical shell and ambient flow is taken into account. In addition, an external source of radiative heat transfer is modeled. Herein, the heat conduction problem is assumed to be heterogeneous which is due to dependence of the conductivity coefficient on temperature. Kirchhoff transformation followed by an integral transform method are used to solve present heterogeneous heat conduction problem. Green's functions are applied to find the solution of final ordinary differential equations. Verification of the present analytical solution is obtained through comparing the analytical results with those of a second order finite difference method. Good agreement between the analytical and numerical results has been found. In order to evaluate the capability of our analytical solution in solving real industrial problems, the temperature distributions in a typical pressure vessel and a pin fin are calculated. [ABSTRACT FROM AUTHOR]

• This proposed model was able to simulate the back-relaxation phenomenon in the IPMC actuators. • This paper provide a more comprehensive governing equation that can be used for vibration analysis and real time control applications. • The proposed model can easily extend to analysis the other types of ionic polymer actuators and sensors. A nonlinear dynamic viscoelastic model of an ionic polymer metal composite (IPMC) actuator is presented based on the viscoelastic constitutive equations and an energy-based variational approach to acquire dynamic and quasi-static response. For this purpose, the equation of motion of an IPMC actuator is obtained by considering the actuator as a cantilever viscoelastic beam and utlizing continuum mechanic relations, Euler-Bernoulli beam theory, Hamilton's principle, and electrochemical properties of ionomer. This new model can simulate time-dependent tip displacement of the IPMC actuators and predict the back-relaxation that occurs in the Nafion-based IPMC actuator under an imposed step voltage. The transfer function derived from the proposed model provides a good frequency response prediction in comparison to the experimental results.This model also can be utilized to obtain mechanical response and vibration analysis of the IPMC sensors. [ABSTRACT FROM AUTHOR]

Maltogenic amylases (MAases, EC 3.2.1.133) have been gotten much attention due to their various applications in industry and commercial processes. MAases belong to subfamily 20 of glycoside hydrolase family 13 (NPase or CDases subfamily) and they have important differences with other members of the family. This enzyme consists of two subunits which form two active sites in the dimer form by binding of the central domain of each subunit to the N domain of the next one (domain-swapping dimeric structure). Allosterism is a possible way of regulating enzymatic activity and no evidence has been found regarding to the cooperativity and correlation between MAases subunits, therefore in this study the allosteric behavior of MAases from a native strain ( Geobacillus sp. Gh6) was investigated. Unlike other members of α-amylase family, MAases showed positive cooperativity between their subunits and the enzyme exhibited sigmoidal nature towards all three cyclodextrin (CD) substrates with a Hill constant ( nH ) value equal to 2, 1.6 and 1.1 for α-CD, β-CD and γ-CD, respectively. On further analysis, the effect of glucose and maltose as MAases allosteric effectors in the presence of β-CD substrate showed that these two effectors had a biphasic effect; while they stimulated the enzyme activity at low concentrations (with a decrease in Hill constant), these metabolites acted as allosteric inhibitors at higher concentrations. Due to the key role of MAases in carbohydrate metabolization, an efficient regulating system for this enzyme is required. In this experiment, for the first time the allosteric properties of MAases were observed and investigated. [ABSTRACT FROM AUTHOR]

Pharmaceutical development of biosimilars is primarily focused on meeting the regulatory requirements for analytical comparability of the product's critical quality attributes (CQAs), concerning safety and efficacy, to those of the originator drug of interest. To this end, the early adoption of a systematic science-based approach, as guided by quality-by-design (QbD) principles, is crucial due to the blind starting point where the same insights of an originator developer into the challenges of a given biopharmaceutical and its manufacturing process are lacking. In this study, we devised a pharmaceutical QbD-guided approach to undertake the biosimilar development of Pembrolizumab (Keytruda®), the ace of therapeutic monoclonal antibodies (mAbs) in terms of approved indications and market sales, and its manufacturing process development. Quality target product profile (QTPP) for Pembrolizumab biosimilar product was assembled using publicly available information on Keytruda®. Upon preliminary analyses of four different lots of Keytruda®, the product CQAs and their acceptable ranges of specification were determined via risk assessment based on the relevant pharmaceutical development quality guidelines, particularly those of the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH). The development and clone selection of Chinese Hamster Ovary (CHO) DG44 cell line was performed using DHFR expression vectors and Methotrexate (MTX) selective pressure. The CHO clone stably expressing relatively higher mAb titer (∼1200 mg/l) in small-scale shake-flask cultures, with the highest similarity of the CQAs charge variants contents (CVCs), N-glycan profile, and biological potency to those of Keytruda® reference standard was selected as the lead clone and the produced Pembrolizumab candidate was named PSG-024. The upstream process (USP) and downstream process (DSP) developments for production were started with the process evaluation screening experiments for the identification of critical process parameters (CPPs) founded upon the prior knowledge on different process stages, input process parameters (iPPs), output process parameters (oPPs), and their impacts on product CQAs. Thereby, screening experiments of USP fed-batch cell culture in 5-liter bioreactor resulted in improvement of PSG-024 expression titer to 2060 ± 70 mg/l and selection of the iPPs feed amount (A), glucose setpoint (B), culture temperature (C), and agitation rate (D) for the optimization design of experiments (DoEs) mainly focused on the CQA acidic CVC and the oPPs mAb expression yield. The USP optimization DoEs using response surface methodology (RSM) yielded valid prediction models and optimal conditions of A = 35%, B = 4.5 g/l, C = 37 °C, and D = 160–220 rpm, which resulted in the final PSG-024 expression titer of 3170 ± 40 mg/l without an excessive rise in acidic CVC. The DSP screening experiments led to achieving the mAb recovery rates of 94% ± 3% and 71.5% ± 3.5% for affinity (capture) and cation-exchange (polishing) chromatography stages, respectively. The capture eluate buffer and viral inactivation conditions were optimized to prevent mAb eluate turbidity and protein aggregation. Moreover, the polishing stage optimization DoEs via one-factor-at-a-time method focused on wash and elution steps for control of the acidic CVC CQA and achieving >80% mAb recovery rate. By shifting to Step elution from the primary salt gradient method and considering an additional intermediate wash step, the maximum mAb recovery of 87% ± 1.5% was achievable while maintaining the CQA acidic CVC within the acceptable range. The consistency of final analytical comparability of PSG-024 demonstrated the effectiveness of the adopted pharmaceutical QbD approach for Pembrolizumab biosimilar development, paving the way for the technology transfer to the client to proceed further development. [Display omitted] [ABSTRACT FROM AUTHOR]