Deep Learning For Automatic Gross Tumor Volumes Contouring in Esophageal Cancer Based on Contrast-Enhanced Computed Tomography Images: A Multi-Institutional Study.

PURPOSE: To develop and externally validate an automatic artificial intelligence (AI) tool for delineating gross tumor volume (GTV) in patients with esophageal squamous cell carcinoma (ESCC), which can assist in neo-adjuvant or radical radiation therapy treatment planning. METHODS AND MATERIALS: In this multi-institutional study, contrast-enhanced CT images from 580 eligible ESCC patients were retrospectively collected. The GTV contours delineated by 2 experts via consensus were used as ground truth. A 3-dimensional deep learning model was developed for GTV contouring in the training cohort and internally and externally validated in 3 validation cohorts. The AI tool was compared against 12 board-certified experts in 25 patients randomly selected from the external validation cohort to evaluate its assistance in improving contouring performance and reducing variation. Contouring performance was measured using dice similarity coefficient (DSC) and average surface distance. Additionally, our previously established radiomics model for predicting pathologic complete response was used to compare AI-generated and ground truth contours, to assess the potential of the AI contouring tool in radiomics analysis. RESULTS: The AI tool demonstrated good GTV contouring performance in multicenter validation cohorts, with median DSC values of 0.865, 0.876, and 0.866 and median average surface distance values of 0.939, 0.789, and 0.875 mm, respectively. Furthermore, the AI tool significantly improved contouring performance for half of 12 board-certified experts (DSC values, 0.794-0.835 vs 0.856-0.881, P = .003-0.048), reduced the intra- and interobserver variations by 37.4% and 55.2%, respectively, and saved contouring time by 77.6%. In the radiomics analysis, 88.7% of radiomic features from ground truth and AI-generated contours demonstrated stable reproducibility, and similar pathologic complete response prediction performance for these contours (P = .430) was observed. CONCLUSIONS: Our AI contouring tool can improve GTV contouring performance and facilitate radiomics analysis in ESCC patients, which indicates its potential for GTV contouring during radiation therapy treatment planning and radiomics studies.

Fe, N-Inducing Interfacial Electron Redistribution in NiCo Spinel on Biomass-Derived Carbon for Bi-functional Oxygen Conversion.

Herein, an interfacial electron redistribution is proposed to boost the activity of carbon-supported spinel NiCo2O4 catalyst toward oxygen conversion via Fe, N-doping strategy. Fe-doping into octahedron induces a redistribution of electrons between Co and Ni atoms on NiCo1.8Fe0.2O4@N-carbon. The increased electron density of Co promotes the coordination of water to Co sites and further dissociation. The generation of proton from water improves the overall activity for the oxygen reduction reaction (ORR). The increased electron density of Ni facilitates the generation of oxygen vacancies. The Ni-VO-Fe structure accelerates the deprotonation of *OOH to improve the activity toward oxygen evolution reaction (OER). N-doping modulates the electron density of carbon to form active sites for the adsorption and protonation of oxygen species. Fir wood-derived carbon endows catalyst with an integral structure to enable outstanding electrocatalytic performance. The NiCo1.8Fe0.2O4@N-carbon express high half-wave potential up to 0.86 V in ORR and low overpotential of 270 mV at 10 mA cm-2 in OER. The zinc-air batteries (ZABs) assembled with the as-prepared catalyst achieve long-term cycle stability (over 2000 cycles) with peak power density (180 mWcm-2). Fe, N-doping strategy drives the catalysis of biomass-derived carbon-based catalysts to the highest level for the oxygen conversion in ZABs.

Computer-aided diagnosis of distal metastasis in non-small cell lung cancer by low-dose CT based radiomics and deep learning signatures.

BACKGROUND: This study aimed to develop and validate radiomics and deep learning (DL) signatures for predicting distal metastasis (DM) of non-small cell lung cancer (NSCLC) in low-dose computed tomography (LDCT). METHODS: Images and clinical data were retrospectively collected for 381 NSCLC patients and prospectively collected for 114 patients at the Fifth Affiliated Hospital of Sun Yat-Sen University. Additionally, we enrolled 179 patients from the Jiangmen Central Hospital to externally validate the signatures. Machine-learning algorithms were employed to develop radiomics signature while the DL signature was developed using neural architecture search. The diagnostic efficiency was primarily quantified with the area under receiver operating characteristic curve (AUC). We interpreted the reasoning process of the radiomics signature and DL signature by radiomics voxel mapping and attention weight tracking. RESULTS: A total of 674 patients with pathologically-confirmed NSCLC were included from two institutions, with 143 of them having DM. The radiomics signature achieved AUCs of 0.885, 0.854, and 0.733 in the internal validation, prospective validation, and external validation while those for DL signature were 0.893, 0.786, and 0.780. The proposed signatures achieved a promising performance in predicting the DM of NSCLC and outperformed the approaches proposed in previous studies. Interpretability analysis revealed that both radiomics and DL signatures could detect the variations among voxels inside tumors, which helped in identifying the DM of NSCLC. CONCLUSIONS: Our study demonstrates the potential of LDCT-based radiomics and DL signatures for predicting DM in NSCLC. These signatures could help improve lung cancer screening regarding further diagnostic tests and treatment strategies.

Recent progress on phytoremediation of urban air pollution.

The rapid growth of population and economy has led to an increase in urban air pollutants, greenhouse gases, energy shortages, environmental degradation, and species extinction, all of which affect ecosystems, biodiversity, and human health. Atmospheric pollution sources are divided into direct and indirect pollutants. Through analysis of the sources of pollutants, the self-functioning of different plants can be utilized to purify the air quality more effectively. Here, we explore the absorption of greenhouse gases and particulate matter in cities as well as the reduction of urban temperatures by plants based on international scientific literature on plant air pollution mitigation, according to the adsorption, dust retention, and transpiration functions of plants. At the same time, it can also reduce the occurrence of extreme weather. It is necessary to select suitable tree species for planting according to different plant functions and environmental needs. In the context of tight urban land use, the combination of vertical greening and urban architecture, through the rational use of plants, has comprehensively addressed urban air pollution. In the future, in urban construction, attention should be paid to the use of heavy plants and the protection and development of green spaces. Our review provides necessary references for future urban planning and research.

Radiomics Nomogram with Added Nodal Features Improves Treatment Response Prediction in Locally Advanced Esophageal Squamous Cell Carcinoma: A Multicenter Study.

OBJECTIVE: We aimed to develop and validate a radiomics nomogram and determine the value of radiomic features from lymph nodes (LNs) for predicting pathological complete response (pCR) to neoadjuvant chemoradiotherapy (NCRT) in patients with locally advanced esophageal squamous cell carcinoma (ESCC). METHODS: In this multicenter retrospective study, eligible participants who had undergone NCRT followed by radical esophagectomy were consecutively recruited. Three radiomics models (modelT, modelLN, and modelTLN) based on tumor and LN features, alone and combined, were developed in the training cohort. The radiomics nomogram was developed by incorporating the prediction value of the radiomics model and clinicoradiological risk factors using multivariate logistic regression, and was evaluated using the receiver operating characteristic curve, validated in two external validation cohorts. RESULTS: Between October 2011 and December 2018, 116 patients were included in the training cohort. Between June 2015 and October 2020, 51 and 27 patients from two independent hospitals were included in validation cohorts 1 and 2, respectively. The radiomics modelTLN performed better than the radiomics modelT for predicting pCR. The radiomics nomogram incorporating the predictive value of the radiomics modelTLN and heterogeneous after NCRT outperformed the clinicoradiological model, with an area under the curve (95% confidence interval) of 0.833 (0.765-0.894) versus 0.764 (0.686-0.833) [p = 0.088, DeLong test], 0.824 (0.718-0.909) versus 0.692 (0.554-0.809) [p = 0.012], and 0.902 (0.794-0.984) versus 0.696 (0.526-0.857) [p = 0.024] in all three cohorts. CONCLUSIONS: Radiomic features from LNs could provide additional value for predicting pCR in ESCC patients, and the radiomics nomogram provided an accurate prediction of pCR, which might aid treatment decision.

Phytosphere purification of urban domestic wastewater.

Industrialization and overpopulation have polluted aquatic environments with significant impacts on human health and wildlife. The main pollutants in urban sewage are nitrogen, phosphorus, heavy metals and organic pollutants, which need to be treated with sewage, and the use of aquatic plants to purify wastewater has high efficiency and low cost. However, the effectiveness and efficiency of phytoremediation are also affected by temperature, pH, microorganisms and other factors. The use of biochar can reduce the cost of wastewater purification, and the combination of biochar and nanotechnology can improve the efficiency of wastewater purification. Some aquatic plants can enrich pollutants in wastewater, so it can be considered to plant these aquatic plants in constructed wetlands to achieve the effect of purifying wastewater. Biochar treatment technology can purify wastewater with high efficiency and low cost, and can be further applied to constructed wetlands. In this paper, the latest research progress of various pollutants in wastewater purification by aquatic plants is reviewed, and the efficient treatment technology of wastewater by biochar is discussed. It provides theoretical basis for phytoremediation of urban sewage pollution in the future.

Development and validation of a deep learning radiomics nomogram for preoperatively differentiating thymic epithelial tumor histologic subtypes.

OBJECTIVES: Using contrast-enhanced computed tomography (CECT) and deep learning technology to develop a deep learning radiomics nomogram (DLRN) to preoperative predict risk status of patients with thymic epithelial tumors (TETs). METHODS: Between October 2008 and May 2020, 257 consecutive patients with surgically and pathologically confirmed TETs were enrolled from three medical centers. We extracted deep learning features from all lesions using a transformer-based convolutional neural network and created a deep learning signature (DLS) using selector operator regression and least absolute shrinkage. The predictive capability of a DLRN incorporating clinical characteristics, subjective CT findings and DLS was evaluated by the area under the curve (AUC) of a receiver operating characteristic curve. RESULTS: To construct a DLS, 25 deep learning features with non-zero coefficients were selected from 116 low-risk TETs (subtypes A, AB, and B1) and 141 high-risk TETs (subtypes B2, B3, and C). The combination of subjective CT features such as infiltration and DLS demonstrated the best performance in differentiating TETs risk status. The AUCs in the training, internal validation, external validation 1 and 2 cohorts were 0.959 (95% confidence interval [CI]: 0.924-0.993), 0.868 (95% CI: 0.765-0.970), 0.846 (95% CI: 0.750-0.942), and 0.846 (95% CI: 0.735-0.957), respectively. The DeLong test and decision in curve analysis revealed that the DLRN was the most predictive and clinically useful model. CONCLUSIONS: The DLRN comprised of CECT-derived DLS and subjective CT findings showed a high performance in predicting risk status of patients with TETs. CLINICAL RELEVANCE STATEMENT: Accurate risk status assessment of thymic epithelial tumors (TETs) may aid in determining whether preoperative neoadjuvant treatment is necessary. A deep learning radiomics nomogram incorporating enhancement CT-based deep learning features, clinical characteristics, and subjective CT findings has the potential to predict the histologic subtypes of TETs, which can facilitate decision-making and personalized therapy in clinical practice. KEY POINTS: • A non-invasive diagnostic method that can predict the pathological risk status may be useful for pretreatment stratification and prognostic evaluation in TET patients. • DLRN demonstrated superior performance in differentiating the risk status of TETs when compared to the deep learning signature, radiomics signature, or clinical model. • The DeLong test and decision in curve analysis revealed that the DLRN was the most predictive and clinically useful in differentiating the risk status of TETs.

Motor imagery electroencephalogram classification algorithm based on joint features in the spatial and frequency domains and instance transfer.

Introduction: Motor imagery electroencephalography (MI-EEG) has significant application value in the field of rehabilitation, and is a research hotspot in the brain-computer interface (BCI) field. Due to the small training sample size of MI-EEG of a single subject and the large individual differences among different subjects, existing classification models have low accuracy and poor generalization ability in MI classification tasks. Methods: To solve this problem, this paper proposes a electroencephalography (EEG) joint feature classification algorithm based on instance transfer and ensemble learning. Firstly, the source domain and target domain data are preprocessed, and then common space mode (CSP) and power spectral density (PSD) are used to extract spatial and frequency domain features respectively, which are combined into EEG joint features. Finally, an ensemble learning algorithm based on kernel mean matching (KMM) and transfer learning adaptive boosting (TrAdaBoost) is used to classify MI-EEG. Results: To validate the effectiveness of the algorithm, this paper compared and analyzed different algorithms on the BCI Competition IV Dataset 2a, and further verified the stability and effectiveness of the algorithm on the BCI Competition IV Dataset 2b. The experimental results show that the algorithm has an average accuracy of 91.5% and 83.7% on Dataset 2a and Dataset 2b, respectively, which is significantly better than other algorithms. Discussion: The statement explains that the algorithm fully exploits EEG signals and enriches EEG features, improves the recognition of the MI signals, and provides a new approach to solving the above problem.

The carotid web: Current research status and imaging features.

The carotid web is commonly found in the carotid bulb or the beginning of the internal carotid artery. It presents as a thin layer of proliferative intimal tissue originating from the arterial wall and extending into the vessel lumen. A large body of research has proven that the carotid web is a risk factor for ischemic stroke. This review summarizes the current research status of the carotid web and focuses on its imaging presentation.

Enhanced Yield of Large-Sized Ti3C2Tx MXene Polymers Nanosheets via Cyclic Ultrasonic-Centrifugal Separation.

Water pollution has spurred the development of membrane separation technology as a potential means of solving the issue. In contrast to the irregular and asymmetric holes that are easily made during the fabrication of organic polymer membranes, forming regular transport channels is essential. This necessitates the use of large-size, two-dimensional materials that can enhance membrane separation performance. However, some limitations regarding yield are associated with preparing large-sized MXene polymer-based nanosheets, which restrict their large-scale application. Here, we propose a combination of wet etching and cyclic ultrasonic-centrifugal separation to meet the needs of the large-scale production of MXene polymers nanosheets. It was found that the yield of large-sized Ti3C2Tx MXene polymers nanosheets reached 71.37%, which was 2.14 times and 1.77 times higher than that prepared with continuous ultrasonication for 10 min and 60 min, respectively. The size of the Ti3C2Tx MXene polymers nanosheets was maintained at the micron level with the help of the cyclic ultrasonic-centrifugal separation technology. In addition, certain advantages of water purification were evident due to the possibility of attaining the pure water flux of 36.5 kg m-2 h-1 bar-1 for the Ti3C2Tx MXene membrane prepared with cyclic ultrasonic-centrifugal separation. This simple method provided a convenient way for the scale-up production of Ti3C2Tx MXene polymers nanosheets.

Phytoremediation of cadmium from soil, air and water.

Potentially toxic elements (PTEs) pose a great threat to ecosystems and long-term exposure causes adverse effects to wildlife and humans. Cadmium induces a variety of diseases including cancer, kidney dysfunction, bone lesions, anemia and hypertension. Here we review the ability of plants to accumulate cadmium from soil, air and water under different environmental conditions, focusing on absorption mechanisms and factors affecting these. Cadmium possess various transport mechanisms and pathways roughly divided into symplast and apoplast pathway. Excessive cadmium concentrations in the environment affects soil properties, pH and microorganism composition and function and thereby plant uptake. At the same time, plants resist cadmium toxicity by antioxidant reaction. The differences in cadmium absorption capacity of plants need more exploration to determine whether it is beneficial for crop breeding or genetic modification. Identify whether plants have the potential to become hyperaccumulator and avoid excessive cadmium uptake by edible plants. The use of activators such as wood vinegar, GLDA (Glutamic acid diacetic acid), or the placement of earthworms and fungi can speed up phytoremediation of plants, thereby reducing uptake of crop varieties and reducing human exposure, thus accelerating food safety and the health of the planet.

Identifying Solitary Granulomatous Nodules from Solid Lung Adenocarcinoma: Exploring Robust Image Features with Cross-Domain Transfer Learning.

PURPOSE: This study aimed to find suitable source domain data in cross-domain transfer learning to extract robust image features. Then, a model was built to preoperatively distinguish lung granulomatous nodules (LGNs) from lung adenocarcinoma (LAC) in solitary pulmonary solid nodules (SPSNs). METHODS: Data from 841 patients with SPSNs from five centres were collected retrospectively. First, adaptive cross-domain transfer learning was used to construct transfer learning signatures (TLS) under different source domain data and conduct a comparative analysis. The Wasserstein distance was used to assess the similarity between the source domain and target domain data in cross-domain transfer learning. Second, a cross-domain transfer learning radiomics model (TLRM) combining the best performing TLS, clinical factors and subjective CT findings was constructed. Finally, the performance of the model was validated through multicentre validation cohorts. RESULTS: Relative to other source domain data, TLS based on lung whole slide images as source domain data (TLS-LW) had the best performance in all validation cohorts (AUC range: 0.8228-0.8984). Meanwhile, the Wasserstein distance of TLS-LW was 1.7108, which was minimal. Finally, TLS-LW, age, spiculated sign and lobulated shape were used to build the TLRM. In all validation cohorts, The AUC ranges were 0.9074-0.9442. Compared with other models, decision curve analysis and integrated discrimination improvement showed that TLRM had better performance. CONCLUSIONS: The TLRM could assist physicians in preoperatively differentiating LGN from LAC in SPSNs. Furthermore, compared with other images, cross-domain transfer learning can extract robust image features when using lung whole slide images as source domain data and has a better effect.

A discussion of microplastics in soil and risks for ecosystems and food chains.

Microplastics are among the major contaminations in terrestrial and marine environments worldwide. These persistent organic contaminants composed of tiny particles are of concern due to their potential hazards to ecosystem and human health. Microplastics accumulates in the ocean and in terrestrial ecosystems, exerting effects on living organisms including microbiomes, fish and plants. While the accumulation and fate of microplastics in marine ecosystems is thoroughly studied, the distribution and biological effects in terrestrial soil call for more research. Here, we review the sources of microplastics and its effects on soil physical and chemical properties, including water holding capacity, bulk density, pH value as well as the potential effects to microorganisms and animals. In addition, we discuss the effects of microplastics in combination with other toxic environmental contaminants including heavy metals and antibiotics on plant growth and physiology, as well as human health and possible degradation and remediation methods. This reflect is an urgent need for monitoring projects that assess the toxicity of microplastics in soil and plants in various soil environments. The prospect of these future research activities should prioritize microplastics in agro-ecosystems, focusing on microbial degradation for remediation purposes of microplastics in the environment.

A strategy for bioremediation of nuclear contaminants in the environment.

Radionuclides released from nuclear contamination harm the environment and human health. Nuclear pollution spread over large areas and the costs associated with decontamination is high. Traditional remediation methods include both chemical and physical, however, these are expensive and unsuitable for large-scale restoration. Bioremediation is the use of plants or microorganisms to remove pollutants from the environment having a lower cost and can be upscaled to eliminate contamination from soil, water and air. It is a cheap, efficient, ecologically, and friendly restoration technology. Here we review the sources of radionuclides, bioremediation methods, mechanisms of plant resistance to radionuclides and the effects on the efficiency of biological adsorption. Uptake of radionuclides by plants can be facilitated by the addition of appropriate chemical accelerators and agronomic management, such as citric acid and intercropping. Future research should accelerate the use of genetic engineering and breeding techniques to screen high-enrichment plants. In addition, field experiments should be carried out to ensure that this technology can be applied to the remediation of nuclear contaminated sites as soon as possible.

Pyrolysis of Aesculus chinensis Bunge Leaves as for Extracted Bio-Oil Material.

Biomass rapid pyrolysis technology is easy to implement in continuous production and industrial application, and has become one of the leading technologies in the field of world renewable energy development. Agricultural and forestry waste is an important resource of renewable energy in China. In general, abandoned leaves in forest areas cause serious waste of resources. Its utilization may help to settle the problems of energy deficiency and environment pollution. In this study, Aesculus chinensis Bunge leaves (A. Bunge) are used as the research object to study the pyrolysis and extract. The results showed that there are a lot of bioactive components in A. Bunge leaves extract, including acetamide, 5-hydroxymethylfurfural, R-limonene, d-mannose, and dihydroxyacetone. The active components of A. Bunge leaves supply scientific evidence for the exploration and exploitation of this plant. The pyrolysis products of A. Bunge leaves are rich in organic acids, aldehydes, and ketones, which means that A. Bunge leaves can be used as a crude material for the manufacturing of bio-oil or bio-fuel. The pyrolysis products include batilol, pregnenolone, benzoic acid, butyrolactone, and propanoic acid, which can be used in biological medicine, chemical crude materials, and industrial raw material reagents. Therefore, A. Bunge leaves can be used as a good crude material for bio-oil or biofuel production. Combining A. Bunge leaves and fast pyrolysis methods can effectively solve the problem of forestry and agricultural residues in the future.

Nano Catalysis of Biofuels and Biochemicals from Cotinus coggygria Scop. Wood for Bio-Oil Raw Material.

Cotinus coggygria Scop. as a precious landscape shrub and a good afforestation species that is used in the pharmaceutical industry. In this paper, TG-FTIR, TG-DTG, and Py-GC/MS were used to study the biomaterials of Cotinus coggygria used as biofuels and biochemicals under the catalysis of nano-Mo/Fe2O3. The wood powder was extracted using a methanol/benzene solution, and the extract was analyzed by FTIR and GC-MS. The results showed that the pyrolysis products of Cotinus coggygria wood were rich in phenols, alcohols, and biofuels. The metal nano-Mo powder played a catalytic role in the interpretation of the gas in the species, where it accelerates gas products. Metal nano-Fe2O3 has a certain flame-retardant effect on the burning process of Cotinus coggygria wood, and the residual amount of pyrolysis is greater. The contents of the extract Formamide, 1-Hexanol, Levodopa, and 9,12-Octadecadienoic acid (Z,Z)- are not only widely used industrially but also play an important role in medicine. Cotinus coggygria is therefore an excellent biomaterial for biofuels and biochemicals.

Predicting lymphovascular invasion in clinically node-negative breast cancer detected by abbreviated magnetic resonance imaging: Transfer learning vs. radiomics.

Objective: To compare the performance of abbreviated breast magnetic resonance imaging (AB-MRI)-based transfer learning (TL) algorithm and radionics analysis for lymphovascular invasion (LVI) prediction in patients with clinically node-negative invasive breast cancer (IBC). Methods: Between November 2017 and October 2020, 233 clinically node-negative IBCs detected by AB-MRI were retrospectively enrolled. One hundred thirty IBCs from center 1 (37 LVI-positive and 93 LVI-negative) were assigned as the training cohort and 103 from center 2 (25 LVI-positive and 78 LVI-negative) as the validation cohort. Based on AB-MRI, a TL signature (TLS) and a radiomics signature (RS) were built with the least absolute shrinkage and selection operator (LASSO) logistic regression. Their diagnostic performances were validated and compared using areas under the receiver operating curve (AUCs), net reclassification improvement (NRI), integrated discrimination improvement (IDI), decision curve analysis (DCA), and stratification analysis. A convolutional filter visualization technique was used to map the response areas of LVI on the AB-MRI. Results: In the validation cohort, compared with RS, the TLS showed better capability in discriminating LVI-positive from LVI-negative lesions (AUC: 0.852 vs. 0.726, p < 0.001; IDI = 0.092, p < 0.001; NRI = 0.554, p < 0.001). The diagnostic performance of TLS was not affected by the menstrual state, molecular subtype, or contrast agent type (all p > 0.05). Moreover, DCA showed that the TLS added more net benefit than RS for clinical utility. Conclusions: An AB-MRI-based TLS was superior to RS for preoperative LVI prediction in patients with clinically node-negative IBC.

Incorporation of Nanocatalysts for the Production of Bio-Oil from Staphylea holocarpa Wood.

Biomass has been recognized as the most common source of renewable energy. In recent years, researchers have paved the way for a search for suitable biomass resources to replace traditional fossil fuel energy and provide high energy output. Although there are plenty of studies of biomass as good biomaterials, there is little detailed information about Staphylea holocarpa wood (S. holocarpa) as a potential bio-oil material. The purpose of this study is to explore the potential of S. holocarpa wood as a bio-oil. Nanocatalyst cobalt (II) oxide (Co3O4) and Nickel (II) oxide (NiO) were used to improve the production of bio-oil from S. holocarpa wood. The preparation of biofuels and the extraction of bioactive drugs were performed by the rapid gasification of nanocatalysts. The result indicated that the abundant chemical components detected in the S. holocarpa wood extract could be used in biomedicine, cosmetics, and biofuels, and have a broad industrial application prospect. In addition, nanocatalyst cobalt tetraoxide (Co3O4) could improve the catalytic cracking of S. holocarpa wood and generate more bioactive molecules at high temperature, which is conducive to the utilization and development of S. holocarpa wood as biomass. This is the first time that S. holocarpa wood was used in combination with nanocatalysts. In the future, nanocatalysts can be used to solve the problem of sustainable development of biological resources.

A brain-like classification method for computed tomography images based on adaptive feature matching dual-source domain heterogeneous transfer learning.

Transfer learning can improve the robustness of deep learning in the case of small samples. However, when the semantic difference between the source domain data and the target domain data is large, transfer learning easily introduces redundant features and leads to negative transfer. According the mechanism of the human brain focusing on effective features while ignoring redundant features in recognition tasks, a brain-like classification method based on adaptive feature matching dual-source domain heterogeneous transfer learning is proposed for the preoperative aided diagnosis of lung granuloma and lung adenocarcinoma for patients with solitary pulmonary solid nodule in the case of small samples. The method includes two parts: (1) feature extraction and (2) feature classification. In the feature extraction part, first, By simulating the feature selection mechanism of the human brain in the process of drawing inferences about other cases from one instance, an adaptive selected-based dual-source domain feature matching network is proposed to determine the matching weight of each pair of feature maps and each pair of convolution layers between the two source networks and the target network, respectively. These two weights can, respectively, adaptive select the features in the source network that are conducive to the learning of the target task, and the destination of feature transfer to improve the robustness of the target network. Meanwhile, a target network based on diverse branch block is proposed, which made the target network have different receptive fields and complex paths to further improve the feature expression ability of the target network. Second, the convolution kernel of the target network is used as the feature extractor to extract features. In the feature classification part, an ensemble classifier based on sparse Bayesian extreme learning machine is proposed that can automatically decide how to combine the output of base classifiers to improve the classification performance. Finally, the experimental results (the AUCs were 0.9542 and 0.9356, respectively) on the data of two center data show that this method can provide a better diagnostic reference for doctors.