Deep Process Research Articles

Life cycle assessment for evaluation of novel solvents and technologies: A case study of flavonoids extraction from Ginkgo biloba leaves

Innovative solvents such as deep eutectic solvents (DESs) and process intensification technologies assisted by ultrasound have been demonstrated to be promising pathways for enhancing solid-liquid extraction. Nevertheless, quantitative and systematic knowledge of their environmental impact is still limited. In this work, a case study of flavonoids extraction from Ginkgo biloba leaves was evaluated by using life cycle assessment (LCA) for comparison of three extraction scenarios. The first used DES as extractant (DESE), and the other two adopted ethanol, including heat reflux extraction (HRE), and ultrasound-assisted extraction (UAE). Among eight key midpoints investigated, all these from UAE were 10.0 %–80.0 % lower than from DESE and HRE except water consumption. The UAE was the eco-friendliest option due to its higher extraction yield, shorter duration and lower solvent consumption. The DESE exhibited the lowest water consumption, the highest freshwater ecotoxicity and human carcinogenic toxicity, while HRE had the highest impacts for the other 6 midpoints. Moreover, solvent production was the key contributor for all the categories. The standardized sensitivity analysis showed that the overall environmental footprint can be further decreased by 15.4 % for DESE pathways via substituting choline chloride/glycerine with choline chloride/ethylene glycol. Furthermore, all pathways using DESs had higher standardized impacts than those employing ethanol from sugarcane or wood. Replacing ethanol from maize with other feedstocks can significantly lessen the overall impacts, among which the UAE using ethanol from sugarcane demonstrated the least environmental impacts. The promotion of DESs as “green and sustainable” alternative to traditional solvents requires careful consideration.

A Hyperspectral Change Detection (HCD-Net) Framework Based on Double Stream Convolutional Neural Networks and an Attention Module

Human activities and natural phenomena continually transform the Earth’s surface, presenting ongoing challenges to the environment. Therefore, the accurate and timely monitoring and prediction of these alterations are essential for devising effective solutions and mitigating environmental impacts in advance. This study introduces a novel framework, called HCD-Net, for detecting changes using bi-temporal hyperspectral images. HCD-Net is built upon a dual-stream deep feature extraction process, complemented by an attention mechanism. The first stream employs 3D convolution layers and 3D Squeeze-and-Excitation (SE) blocks to extract deep features, while the second stream utilizes 2D convolution and 2D SE blocks for the same purpose. The deep features from both streams are then concatenated and processed through dense layers for decision-making. The performance of HCD-Net is evaluated against existing state-of-the-art change detection methods. For this purpose, the bi-temporal Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) hyperspectral dataset was utilized to assess the change detection performance. The findings indicate that HCD-Net achieves superior accuracy and the lowest false alarm rate among the compared methods, with an overall classification accuracy exceeding 96%, and a kappa coefficient greater than 0.9.

Wet cleaning of Ta-based extreme ultraviolet photomasks at room temperature

Owing to the wavelength-dependent limits of the deep ultraviolet exposure process, the semiconductor industry introduced extreme ultraviolet (EUV) lithography operating at a 13.5 nm wavelength. Traditional photomasks employ pellicles for protection; however, EUV-specific pellicles are not widely applicable to commercial processes, requiring the development of a EUV photomask cleaning method. In this study, a wet cleaning method for Ta-based EUV photomasks at room temperature was systematically examined in terms of key parameters, including the pattern step height, surface topography, and particulate count, via atomic force microscopy and x-ray reflectivity. Post sulfuric acid-hydrogen peroxide mixture (SPM) treatment, the photomask exhibited a stable step height, indicating minimal pattern degradation. Additionally, discernible alterations in the surface roughness and a decrease in particle count were observed, further indicating to the effectiveness of SPM-mediated cleaning. Conversely, following standard clean-1 (SC-1) treatment, while the pattern step height remained relatively unchanged, a notable increase in surface irregularities and macroscopic particulates was observed, suggesting a suboptimal cleaning efficiency of the SC-1 solution despite its potential for pattern structure preservation. Our room temperature wet cleaning method efficiently reduces wear-out and successfully eliminates contaminants, potentially prolonging the EUV photomask’s productivity and durability.

Evolution learning method to derive missing elements and optimal classification under the connection hesitant fuzzy environment

Generally, in the qualitative decision-making process, the attributes are independent of each other, thus the given attributes' evaluation data could ignore their connection relationships and make a decision based on incomplete information. To do this, this paper first introduces the hesitant fuzzy set (HFS), an emerging evaluation tool, to develop the connection HFS (CHFS). Furthermore, considering the missing-element situation, which is a normal phenomenon in the experts' subjective evaluation process, we extend the CHFS to the missing CHFS (M−CHFS). Thus, the M−CHFS can quantitatively express the connection relationship and missing-element characteristic simultaneously. Then, to derive the missing elements in the M−CHFS evaluation information, this paper proposes the hesitant fuzzy generative adversarial network to convert the M−CHFS to the CHFS according to a deep learning process. After that, we propose a hesitant fuzzy long short-term memory network to complete the evolution learning on the obtained CHFS and then get the classification results, occurrence probabilities, and optimal classification. Then, this paper develops an evolution learning algorithm introducing the decision makers' subjective guidance and applies it to an example of intelligent manufacturing analysis. The calculation process and derived results present the feasibility and effectiveness of the given methods in this paper.

Petrogenesis of Late Mesozoic trachydacites in the Jiaolai Basin, Eastern China: Implications for the destruction of the North China Craton

The North China Craton has witnessed a large-scale lithospheric thinning event in the Mesozoic. The volcanic rocks of the Qingshan Group formed during the destruction of the North China Craton, therefore retaining important information to the deep processes in this region. Zircon U-Pb dating and geochemical analyses have been carried out on volcanic rocks of the Qingshan Group collected from southeastern margin of the Jiaolai Basin. In combination with previously published data on the volcanic rocks of the Qingshan Group, the newly acquired data show that: (1) the volcanic rocks of Qingshan Group in Mashan have SiO2 content of 62.82 ∼ 66.3 wt% and total alkali content (K2O + Na2O) of 7.64 ∼ 9.34 wt%. According to TAS diagram, they belong to trachydacite which experienced strong fractionation between LREEs and HREEs with depletion of HFSEs and enrichment of LILEs. (2) The crystallization age of trachydacite in Mashan is 116.4 ± 2.3 Ma (MSWD = 1.9, n = 9). The magma originated from the partial melting of the ancient lower crust and the enriched lithospheric mantle metasomatized by the fluids/melts released from the subducted slab of the Yangtze Craton. (3) The formation age of magmatic rocks in western Shandong is generally earlier than that in eastern Shandong, while the lower crust delamination in western Shandong preceded that in eastern Shandong. The occurrence of magmatism in eastern Shandong is probably attributed to the delamination process in western Shandong as well as the ongoing tectonic activities along the Tan-Lu fault zone. The lithospheric thinning of the North China Craton was mainly controlled by the change of the subduction angle of the Paleo-Pacific plate, and is probably influenced by the continuous activity of the Tan-Lu fault zone in eastern Shandong.

Review of Asymmetric Seafloor Spreading and Oceanic Ridge Jumps in the South China Sea

Seafloor spreading is an important cornerstone of the theory of plate tectonics. Asymmetric seafloor spreading and oceanic ridge jumps are common phenomena in this process and play important roles in controlling oceanic crust accretion, regional tectonics and geological geometric boundaries. As the largest marginal sea in the western Pacific, the South China Sea is an ideal laboratory for dissecting the Wilson cycle of small marginal sea-type ocean basins restricted by surrounding blocks and exploring the deep dynamic processes of confined small ocean basins. In recent years, a lot of research has been conducted on the spreading history of the South China Sea and has achieved fruitful results. However, the detailed dynamic mechanisms of asymmetric seafloor spreading and ridge jumps are still unclear. Therefore, this paper summarizes the basic understanding about the dynamic mechanisms of global asymmetric seafloor spreading and ridge jumps and reviews the related research results of asymmetric seafloor spreading and ridge jumps in the South China Sea. Previous studies have basically confirmed that seafloor spreading in the South China Sea started between ~32 and 34 Ma in the east sub-basin and ended at ~15 Ma in the northwest sub-basin, with at least once oceanic ridge jump in the east sub-basin. The current research mainly focuses on the age of the seafloor spreading in the South China Sea and the location, time and stage of the ridge jumps, but there are relatively few studies on high-resolution lithospheric structure across these ridges and the dynamic mechanism of oceanic ridge jumps. Based on the current research progress, we propose that further studies should focus on the lithosphere–asthenosphere scale in the future, suggesting that marine magnetotelluric and Ocean Bottom Seismometer (OBS) surveys should be conducted across the residual oceanic ridges to perform a detailed analysis of the tectonics magmatism in the east sub-basin to gain insights into the dynamic mechanisms of oceanic ridge jumps and asymmetric seafloor spreading, which can promote understanding of the tectonic evolution of the South China Sea and improve the classical plate tectonics theory that was constructed based on the open ocean basins.

Revolutionizing Robotic Depalletizing: AI-Enhanced Parcel Detecting with Adaptive 3D Machine Vision and RGB-D Imaging for Automated Unloading.

Detecting parcels accurately and efficiently has always been a challenging task when unloading from trucks onto conveyor belts because of the diverse and complex ways in which parcels are stacked. Conventional methods struggle to quickly and accurately classify the various shapes and surface patterns of unordered parcels. In this paper, we propose a parcel-picking surface detection method based on deep learning and image processing for the efficient unloading of diverse and unordered parcels. Our goal is to develop a systematic image processing algorithm that emphasises the boundaries of parcels regardless of their shape, pattern, or layout. The core of the algorithm is the utilisation of RGB-D technology for detecting the primary boundary lines regardless of obstacles such as adhesive labels, tapes, or parcel surface patterns. For cases where detecting the boundary lines is difficult owing to narrow gaps between parcels, we propose using deep learning-based boundary line detection through the You Only Look at Coefficients (YOLACT) model. Using image segmentation techniques, the algorithm efficiently predicts boundary lines, enabling the accurate detection of irregularly sized parcels with complex surface patterns. Furthermore, even for rotated parcels, we can extract their edges through complex mathematical operations using the depth values of the specified position, enabling the detection of the wider surfaces of the rotated parcels. Finally, we validate the accuracy and real-time performance of our proposed method through various case studies, achieving mAP (50) values of 93.8% and 90.8% for randomly sized and rotationally covered boxes with diverse colours and patterns, respectively.

Contactless weighing method based on deep learning and acoustic levitation

Abstract Acoustic weighing is a promising contactless method for screening the mass of micro-nano objects as it avoids contact contamination and losses. Existing acoustic weighing methods determine the mass of an object by detecting its oscillation trajectory with a laser sensor. However, this method suffers from several limitations, such as short measurement distance, poor accuracy in measuring transparent objects, and inducing damage to photosensitive samples. To solve these issues, this work proposes a contactless weighing method based on location-aware neural network (LANet) and acoustic levitation. The proposed LANet is a deep learning-based image processing method that detects object bit oscillation trajectories completely contactless, regardless of the color, shape, and oscillation distance of the levitated object. We employ a cross-stage aggregation module and cross-mixed feature pyramid strategy to build LANet network depth for enhanced feature extraction. In addition, to create a contactless environment, we built an acoustic levitation system, which drives the oscillation of objects. Finally, we verified the accuracy and effectiveness of the method. The results show that the proposed network can accurately detect the oscillation trajectories of various objects with high detection performance, even for small objects in low-contrast backgrounds. Meanwhile, the proposed method can accurately measure the mass of objects with a percentage error of no more than 7.83%.

License plate text recognition using deep learning, NLP, and image processing techniques

Detecting license plates has never been easy, particularly with the proliferation of sophisticated radars on highways and roads. By 2021, the gendarmerie and National Security Road control agents will have access to more than 1 billion smart traffic radars worldwide. This research presents a revolutionary technique for detecting and recognizing Arabic and Latin license plates. After assembling the gathered images to create a novel dataset, we utilized YOLO v7 to locate and identify the number plate in the image as the first step of the suggested procedure. Before the dataset was fed to the detection system, it was manually labeled. Afterward, we improved the recognized license plate using machine learning methods. To do this, we used kernel methods as well as thresholding to get rid of the extra vertical lines on the plate. After that, we employed Arabic OCR along with Easy OCR methods to decipher the Latin and Arabic characters on the number plate. Eventually, the proposed method achieved an F1 score of 98%,with a precision and recall of 97% and 98%, respectively. We also obtained an accuracy of 99% for image segmentation. The segmentation and detection results from the suggested strategy have shown satisfactory results.

Automatic segmentation of lamellar eutectoids in iron oxide scales using optimized U-net network

The variation in lamellar spacing of eutectoids in iron oxide affects the surface quality of the hot-rolled strip and its subsequent deep processing. In order to investigate the lamellar spacing numerically, an accurate segmentation of Fe and Fe3O4 phases is required in the structural picture. In this paper, a dataset was generated for identifying Fe and Fe3O4 in eutectoids. Also, a squeeze-and-excitation network module was integrated for improving the accuracy in image segmentation. The obtained experimental results show that the PA and mIoU could reach 96.83% and 94.49% in optimized U-net, indicating its excellent segmentation capability compared with traditional image segmentation networks. Using this method, fast and accurate extraction of the characteristic information about the picture can be realized, which would help in investigating the phase transition behavior of the eutectoid transformation in the Fe–O system.

Role of Cyclins and Cytoskeletal Proteins in Endometriosis: Insights into Pathophysiology.

Endometriosis is a gynecological condition where endometrium-like tissue grows outside the uterus, posing challenges in understanding and treatment. This article delves into the deep cellular and molecular processes underlying endometriosis, with a focus on the crucial roles played by cyclins and cytoskeletal proteins in its pathogenesis, particularly in the context of Epithelial-Mesenchymal Transition (EMT). The investigation begins by examining the activities of cyclins, elucidating their diverse biological roles such as cell cycle control, proliferation, evasion of apoptosis, and angiogenesis among ectopic endometrial cells. A comprehensive analysis of cytoskeletal proteins follows, emphasizing their fundamental biological roles and their specific significance to endometriotic cell features. This review sheds light on the interconnected pathways through which cyclins and cytoskeletal proteins converge, contributing to the genesis and progression of endometriosis. Understanding these molecular complexities not only provides insight into the underlying causes of the disease but also holds promise for the development of specific therapeutic approaches, ushering in a new era in the management of this devastating disorder.

Deepfakes: current and future trends

Advances in Deep Learning (DL), Big Data and image processing have facilitated online disinformation spreading through Deepfakes. This entails severe threats including public opinion manipulation, geopolitical tensions, chaos in financial markets, scams, defamation and identity theft among others. Therefore, it is imperative to develop techniques to prevent, detect, and stop the spreading of deepfake content. Along these lines, the goal of this paper is to present a big picture perspective of the deepfake paradigm, by reviewing current and future trends. First, a compact summary of DL techniques used for deepfakes is presented. Then, a review of the fight between generation and detection techniques is elaborated. Moreover, we delve into the potential that new technologies, such as distributed ledgers and blockchain, can offer with regard to cybersecurity and the fight against digital deception. Two scenarios of application, including online social networks engineering attacks and Internet of Things, are reviewed where main insights and open challenges are tackled. Finally, future trends and research lines are discussed, pointing out potential key agents and technologies.

Cyclical interplay between L2 writing, WCF processing and rewriting: Explaining modification and consolidation in L2 development

Although different empirical studies have recently explored the L2 writing–SLA interfaces, how L2 writers modify and consolidate their linguistic knowledge with different levels of perceived complexity remains underexplored. Featuring the qualitative phase of a longitudinal, mixed-methods study with the aim of complementarity, the current study explored the perspectives of 20 low- to mid-intermediate L2 writers about how their consistent engagement in an iterative process of writing, feedback processing and rewriting contributed to L2 development. In-depth explanations and numerous script-specific exemplifications in semi-structured interviews showcased the language learning potential of acting on direct WCF on an extended number of grammatical features over time, resulting in deep WCF processing and substantial cognitive and affective engagement with feedback. Results suggested participants' reliance on the prolonged process to (re)formulate and confirm hypotheses and increasingly accumulate category-based WCF instances from previous and subsequent compositions to abstract most of the underlying rules embedded in direct WCF. The participants explained how different challenges against rule abstraction were resolved, and why some remained unresolved. The study proposes a model illustrating a synergistic interaction between writing, feedback processing and rewriting for L2 writers to gradually consolidate linguistic knowledge. Theoretical, methodological and pedagogical implications of the study are discussed.

A Novel Fully Convolutional Auto-Encoder Based on Dual Clustering and Latent Feature Adversarial Consistency for Hyperspectral Anomaly Detection

With the development of artificial intelligence, the ability to capture the background characteristics of hyperspectral imagery (HSI) has improved, showing promising performance in hyperspectral anomaly detection (HAD) tasks. However, existing methods proposed in recent years still suffer from certain limitations: (1) Constraints are lacking in the deep feature learning process in terms of the issue of the absence of prior background and anomaly information. (2) Hyperspectral anomaly detectors with traditional self-supervised deep learning methods fail to ensure prioritized reconstruction of the background. (3) The architecture of fully connected deep networks in hyperspectral anomaly detectors leads to low utilization of spatial information and the destruction of the original spatial relationship in hyperspectral imagery and disregards the spectral correlation between adjacent pixels. (4) Hypotheses or assumptions for background and anomaly distributions restrict the performance of many hyperspectral anomaly detectors because the distributions of background land covers are usually complex and not assumable in real-world hyperspectral imagery. In consideration of the above problems, in this paper, we propose a novel fully convolutional auto-encoder based on dual clustering and latent feature adversarial consistency (FCAE-DCAC) for HAD, which is carried out with self-supervised learning-based processing. Firstly, density-based spatial clustering of applications with a noise algorithm and connected component analysis are utilized for successive spectral and spatial clustering to obtain more precise prior background and anomaly information, which facilitates the separation between background and anomaly samples during the training of our method. Subsequently, a novel fully convolutional auto-encoder (FCAE) integrated with a spatial–spectral joint attention (SSJA) mechanism is proposed to enhance the utilization of spatial information and augment feature expression. In addition, a latent feature adversarial consistency network with the ability to learn actual background distribution in hyperspectral imagery is proposed to achieve pure background reconstruction. Finally, a triplet loss is introduced to enhance the separability between background and anomaly, and the reconstruction residual serves as the anomaly detection result. We evaluate the proposed method based on seven groups of real-world hyperspectral datasets, and the experimental results confirm the effectiveness and superior performance of the proposed method versus nine state-of-the-art methods.

Antibody-Drug Conjugate Overview: a State-of-the-art Manufacturing Process and Control Strategy.

Antibody-drug conjugates (ADCs) comprise an antibody, linker, and drug, which direct their highly potent small molecule drugs to target tumor cells via specific binding between the antibody and surface antigens. The antibody, linker, and drug should be properly designed or selected to achieve the desired efficacy while minimizing off-target toxicity. With a unique and complex structure, there is inherent heterogeneity introduced by product-related variations and the manufacturing process. Here this review primarily covers recent key advances in ADC history, clinical development status, molecule design, manufacturing processes, and quality control.The manufacturing process, especially the conjugation process, should be carefully developed, characterized, validated, and controlled throughout its lifecycle. Quality control is another key element to ensure product quality and patient safety. A patient-centric strategy has been well recognized and adopted by the pharmaceutical industry for therapeutic proteins, and has been successfully implemented for ADCs as well, to ensure that ADC products maintain their quality until the end of their shelf life. Deep product understanding and process knowledge defines attribute testing strategies (ATS). Quality by design (QbD) is a powerful approach for process and product development, and for defining an overall control strategy. Finally, we summarize the current challenges on ADC development and provide some perspectives that may help to give related directions and trigger more cross-functional research to surmount those challenges.

Deep reinforcement learning-assisted extended state observer for run-to-run control in the semiconductor manufacturing process

In the semiconductor manufacturing process, extended state observer (ESO)-based run-to-run (RtR) control is an intriguing solution. Although an ESO-RtR control strategy can effectively compensate for the lumped disturbance, appropriate gains are required. In this article, a cutting-edge deep reinforcement learning (DRL) technique is integrated into ESO-RtR, and a composite control framework of DRL-ESO-RtR is developed. In particular, the well-trained DRL agent serves as an assisted controller, which produces appropriate gains of ESO. The optimized ESO then presents a preferable control recipe for the manufacturing process. Under the RtR framework, the gain adjustment problem of ESO is formulated as a Markov decision process. An efficient state space and reward function are wisely designed using the system’s observable information. Correspondingly, the gain of the ESO is adaptively adjusted to cope with changing environmental disturbances. Finally, a twin-delayed deep deterministic policy gradient algorithm is employed to implement the suggested scheme. The feasibility and superiority of the developed method are validated in a deep reactive ion etching process. Comparative results demonstrate that the presented scheme outperforms the ordinary ESO-RtR controller in terms of disturbance rejection.

Evaluation of vegetable oil blends properties during deep frying process and their influence on the quality of fried potatoes

Using different vegetable oils, it is possible to obtain blends that are less susceptible to oxidation, improving stability. In this work, the objective was evaluate alterations that occurred with two blends developed using soybean, cottonseed, and palm oils: Blend 1 in the proportions of 20:20:60 and Blend 2, 20:60:20, respectively. Batch frying study was conducted under temperature condition of 185°C for 20 hours. In the oils, the acid index, viscosity, color, composition in fatty acids and polar compounds were determined; and in the fried potato samples, moisture content and total lipids were analyzed. Results show that, in oils, there is an increase in the amount of saturated fatty acids greater for Blend 2 by 12.43% at the end of 20 hours of frying. There was an increase in viscosity during frying and there was no significant difference between the acidity of the blends. Both blends have suffered color change due to oil degradation. Oils with more saturated fatty acids, such as Blend 1, present greater stability during potato frying, however, there is a greater incorporation of oil in the finished product. Blend 1 was considered more stable, as it had the formation of 5.3% less polar compounds, as it has a higher amount of palm oil.

Responses of deep soil moisture to direct rainfall and groundwater in the red soil critical zone: A four-stage pattern

Deep soil moisture (DSM) plays a critical role in vadose zone hydrological processes. However, the recharge processes and their underlying mechanisms remain unclear due to the scarcity of comprehensive deep hydrological data, especially in regions characterized by active interactions between surface and subsurface hydrological processes. To address this knowledge gap, we performed an event-based study using a four-year hydrological monitoring dataset, encompassing rainfall, profile soil moisture, and groundwater along a hillslope at the Red Soil Critical Zone Observatory (CZO) in China. A total of 226 rainfall events were classified into four distinct types (I-IV) using a k-means clustering algorithm. The changes in DSM and groundwater table (GWT) were dependent on the rainfall type. The GWT (0.57–0.97) played a greater role in contributing to the DSM than the direct rainfall (0.48–0.64) based on a structural equation model, which was more pronounced from upper slope to foot slope. DSM recharge pattern can be described by a four-stage conceptual model, as indicated by the relationship between DSM and GWT: a slow and linear increase (stage 1), a small plateau (stage 2), a rapid linear rise (stage 3), and finally a great plateau (stage 4). DSM at the small and great plateau were closely related with the field capacity and saturation, respectively. Rainfall type, vadose zone features, and soil physical properties jointly shaped DSM recharge patterns. These findings can offer valuable insights into understanding the deep hydrological processes of the red soil region as well as regions with similar hydroclimatic conditions.

Centralized Database Access: Transformer Framework and LLM/Chatbot Integration-Based Hybrid Model

This paper addresses the challenges associated with the centralized storage of educational materials in the context of a fragmented and disparate database. In response to the increasing demands of modern education, efficient and accessible retrieval of materials for educators and students is essential. This paper presents a hybrid model based on the transformer framework and utilizing an API for an existing large language model (LLM)/chatbot. This integration ensures precise responses drawn from a comprehensive educational materials database. The model architecture uses mathematically defined algorithms for precise functions that enable deep text processing through advanced word embedding methods. This approach improves accuracy in natural language processing and ensures both high efficiency and adaptability. Therefore, this paper not only provides a technical solution to a prevalent problem but also highlights the potential for the continued development and integration of emerging technologies in education. The aim is to create a more efficient, transparent, and accessible educational environment. The importance of this research lies in its ability to streamline material access, benefiting the global scientific community and contributing to the continuous advancement of educational technology.

Sucrose-phosphate osmotic system improves the quality characteristics of reduced-salt salted egg yolk: Profiling from protein structure and lipid distribution perspective

This paper was dedicated to the study of the effect of sucrose-phosphate on aspects of physicochemical properties, lipid distribution and protein structure during the picklig of reduced‐salt salted egg yolk (SEY). This work constructed a reduced‐salt pickling system from a new perspective (promoting osmosis) by using a sucrose-phosphate-salt. Results showed that SEY-28d achieved a desirable salt content (1.07 %), hardness (573.46 g) and springiness (0.65 g). The matured SEY was in excellent quality with orange-red color and loose sandy texture. This was because the lipoprotein aggregated with each other through hydrophobic interaction to form a stable network structure. In addition, the hypertonic environment accelerated salt penetration. These also created good condition for lipid spillage. The results of confocal laser scanning microscope also verified this phenomenon. This work provides important guidance for new reduced‐salt curing of traditional pickled foods, deep processing of SEY, and industry development in the field of poultry egg.