Correction Output Research Articles

Efficient microplastic identification by hyperspectral imaging: A comparative study of spatial resolutions, spectral ranges and classification models to define an optimal analytical protocol

Microplastics (MPs) pollution is a global and challenging issue, necessitating the development of efficient analytical strategies for their detection to monitor their environmental impact.This study aims to define an optimal analytical protocol for characterizing MPs by hyperspectral imaging (HSI), comparing different setups based on spatial resolution, spectral range and classification models. The investigated MPs include polymers commonly found in the environment, such as polystyrene (PS), polypropylene (PP) and high-density polyethylene (HDPE), subdivided in three size classes (1000–2000 μm, 500–1000 μm, 250–500 μm). Furthermore, MP particles with diameters ranging from 30 to 250 μm were assessed to determine the limit of detection (LOD) in the different configurations. Hyperspectral images were acquired with two spatial resolutions, 150 and 30 μm/pixel, and two spectral ranges, 1000–1700 nm (NIR) and 1000–2500 nm (SWIR). Three classification models, Partial Least Square-Discriminant Analysis (PLS-DA), Error Correction Output Coding-Support Vector Machine (ECOC-SVM) and Neural Network Pattern Recognition (NNPR) were tested on the acquired images. The correctness of these models was evaluated by prediction maps and statistical parameters (Recall, Specificity and Accuracy). The results demonstrated that for MP particles larger than 250 μm, the optimal setup is a spatial resolution of 150 μm/pixel and a spectral range of 1000–1700 nm, utilizing a linear classification model like PLS-DA. This approach offers accurate predictions while being time- and cost-efficient. For MPs smaller than 250 μm, a higher spatial resolution of 30 μm/pixel with a spectral range of 1000–2500 nm and a non-linear classification method like ECOC-SVM is preferable. The LOD is 250 μm for the 150 μm/pixel resolution and ranges from 100 to 200 μm for the 30 μm/pixel resolution. These findings provide a valuable guide for selecting the appropriate HSI acquisition conditions and data processing methods to optimally characterize MPs of different sizes.

Sports activity (SA) recognition based on error correcting output codes (ECOC) and convolutional neural network (CNN)

The increasing use of motion sensors is causing major changes in the process of monitoring people's activities. One of the main applications of these sensors is the detection of sports activities, for example, they can be used to monitor the condition of athletes or analyze the quality of sports training. Although the existing sensor-based activity recognition systems can recognize basic activities such as: walking, running, or sitting; they don't perform well in recognizing different types of sports activities. This article introduces a new model based on machine learning (ML) techniques to more accurately distinguish between sports and everyday activities. In the proposed method, the necessary data to detect the type of activity is collected through his two sensors: an accelerometer and a gyroscope attached to a person's foot. For this purpose, the input signals are first preprocessed and then short-time Fourier transform (STFT) is used to describe the characteristics of each signal. In the next step, each STFT matrix is used as input to a convolutional neural network (CNN). This CNN describes various motion characteristics of the sensor in the form of vectors. Finally, a classification model based on error correction output code (ECOC) is used to classify the extracted features and detect the type of SA. The performance of the proposed AS recognition method is evaluated using the DSADS database and the results are compared with previous methods. Based on the results, the proposed method can recognize sports activities with an accuracy of 99.71. Furthermore, the performance of the proposed method based on precision and recall criteria are 99.72 and 99.71, respectively, which are better than the compared methods.

An in silico study of the effects of cardiovascular aging on carotid flow waveforms and indexes in a virtual population.

Cardiovascular aging is strongly associated with increased risk of cardiovascular disease and mortality. Moreover, health and lifestyle factors may accelerate age-induced alterations, such as increased arterial stiffness and wall dilation, beyond chronological age, making the clinical assessment of cardiovascular aging an important prompt for preventative action. Carotid flow waveforms contain information about age-dependent cardiovascular properties, and their ease of measurement via noninvasive Doppler ultrasound (US) makes their analysis a promising tool for the routine assessment of cardiovascular aging. In this work, the impact of different aging processes on carotid waveform morphology and derived indexes is studied in silico, with the aim of establishing the clinical potential of a carotid US-based assessment of cardiovascular aging. One-dimensional (1-D) hemodynamic modeling was employed to generate an age-specific virtual population (VP) of N = 5,160 realistic carotid hemodynamic waveforms. The resulting VP was statistically validated against in vivo aging trends in waveforms and indexes from the literature, and simulated waveforms were studied in relation to age and underlying cardiovascular parameters. In our study, the carotid flow augmentation index (FAI) significantly increased with age (with a median increase of 50% from the youngest to the oldest age group) and was strongly correlated to local arterial stiffening (r = 0.94). The carotid pulsatility index (PI), which showed less pronounced age variation, was inversely correlated with the reflection coefficient at the carotid branching (r = -0.88) and directly correlated with carotid net forward wave energy (r = 0.90), corroborating previous literature where it was linked to increased risk of cerebrovascular damage in the elderly. There was a high correlation between corrected carotid flow time (ccFT) and cardiac output (CO) (r = 0.99), which was not affected by vascular age. This study highlights the potential of carotid waveforms as a valuable tool for the assessment of cardiovascular aging.NEW & NOTEWORTHY An age-specific virtual population was generated based on a 1-D model of the arterial circulation, including newly defined literature-based specific age variations in carotid vessel properties. Simulated carotid flow/velocity waveforms, indexes, and age trends were statistically validated against in vivo data from the literature. A comprehensive study of the impact of aging on carotid flow waveform morphology was performed, and the mechanisms influencing different carotid indexes were elucidated. Notably, flow augmentation index (FAI) was found to be a strong indicator of local carotid stiffness.

Dynamic decoding and dual synthetic data for automatic correction of grammar in low-resource scenario.

Grammar error correction systems are pivotal in the field of natural language processing (NLP), with a primary focus on identifying and correcting the grammatical integrity of written text. This is crucial for both language learning and formal communication. Recently, neural machine translation (NMT) has emerged as a promising approach in high demand. However, this approach faces significant challenges, particularly the scarcity of training data and the complexity of grammar error correction (GEC), especially for low-resource languages such as Indonesian. To address these challenges, we propose InSpelPoS, a confusion method that combines two synthetic data generation methods: the Inverted Spellchecker and Patterns+POS. Furthermore, we introduce an adapted seq2seq framework equipped with a dynamic decoding method and state-of-the-art Transformer-based neural language models to enhance the accuracy and efficiency of GEC. The dynamic decoding method is capable of navigating the complexities of GEC and correcting a wide range of errors, including contextual and grammatical errors. The proposed model leverages the contextual information of words and sentences to generate a corrected output. To assess the effectiveness of our proposed framework, we conducted experiments using synthetic data and compared its performance with existing GEC systems. The results demonstrate a significant improvement in the accuracy of Indonesian GEC compared to existing methods.

Carotid Flow Time Compared with Invasive Monitoring as a Predictor of Volume Responsiveness in ICU patients.

Objectives: Identifying patients who will have an increase in their cardiac output from volume administration is difficult to identify. We propose the use of carotid flow time, which is a non-invasive means to determine if a patient is volume responsive. Methods: Patients admitted to a critical care unit with a pulmonary artery catheter in place were enrolled. We perform a carotid flow time and pulmonary artery catheter measurement of cardiac output pre and post-passive leg raise and comparing the two. An increase of 10% change in the pre- vs. post-passive leg raise measurement would be indicative of a patient who is volume responsive. Results: We identified 8 patients who were volume responsive as determined by the gold standard pulmonary artery catheter. The sensitivity 87.5% and specificity 90.9%. Pearson correlation coefficient between PA-CO measurements and CFT was r=0.8316, indicative of strong correlation between the two measurements. Conclusion: In our patient sample of critically ill patients with pulmonary artery catheters, we found a strong correlation between corrected carotid flow times and cardiac output measurements from pulmonary artery catheters.

Front-End Circuit for Photomultiplier Tube Signal Readout Based on Recognition of Traffic Signal Images

Photoelectric sensing technology plays a crucial role in vehicular equipment, which is equipped with various photoelectric devices to perceive the surrounding environment and avoid traffic lights and vehicles. This research selects the Hamamatsu H9500, a 256-channel, position-sensitive photomultiplier tube, as the test unit. It aims to simplify signal readout while improving the spatial resolution of the photodetector. This research focuses on designing a charge distribution circuit named Discretized Positioning Circuit (DPC) for the photomultiplier tube, with an additional charge-sensitive front-end amplification and shaping circuit. This circuit can convert the weak current signals from the H9500 into voltage signals. The shaping part of the circuit employs an active CR-RC circuit with weak signal amplification capabilities. This circuit is deployed within the photomultiplier tube, strategically positioned on vehicles to recognize various traffic sign images. The front-end shaping circuit is tested in the experiments, which converts square wave voltage into pulse current using a capacitor. It is observed that the current signal has a certain width and the voltage waveform of the CR differential circuit can be obtained by increasing the input impedance to 1 MΩ. During input voltage amplitude testing, the corrected output signal voltage shows a good linear relationship with the input square wave voltage. This designed front-end shaping circuit is used for signal readout in photomultiplier tubes and deployed in vehicular equipment to collect image information of traffic signs. After image processing, satisfactory recognition results are achieved.

Research on the application of improved V-detector algorithm in network intrusion detection

Abstract Network intrusion detection has been widely discussed and studied as an important part of protecting network security. Therefore, this paper presents an in-depth study of the application of an improved V-detector algorithm in network intrusion detection. In this paper, we construct a V-detector intrusion detection model, adopt the “self-oriented” identification principle, and randomly generate detectors with large differences from the health library. A smaller number of detectors are used to compare the data information generated by the computer, and if they are similar, they are judged as intrusions. Intrusion detection experiments are performed on multiple types of networks by using classifiers to determine whether the access to be detected is an attack access. The experimental results show that the model has the lowest false alarm rate for mixed feature networks, with a false alarm rate of only 13% and a detection rate of 89%, with a sample size of 25,987. After the improvement of the V-detector intrusion detection model, the error correction output problem leads to a network intrusion with a miss rate of only 11% and a protection rate of 85%. The experimental data proved that the model has the advantages of large data size and comprehensive intrusion attack types.

Multi-objective optimization of two-stage AC–DC power supply for reliability and efficiency using NSGA-II and meta-heuristic honey bee algorithms

This paper introduces a novel approach to simultaneously optimize the reliability and efficiency of a switching power supply (SPS) using the Non-dominated Sorting Genetic Algorithm (NSGA-II) and Honey Bee Algorithm (HBA). The proposed method uses a Markov model to evaluate the reliability of the converter and maximizes objective functions, including efficiency and reliability, using optimization algorithms to identify the optimum values of the converter parameters. The optimization algorithm determines the optimal values for converter parameters, such as switching frequency, transformer turn ratio, and DC-link voltage. It also fine-tunes circuit components, such as the power factor correction inductor, DC-link capacitor, and output filter components. A sensitivity analysis is carried out to assess the impact of various parameters on the multi-objective functions of the converter. The converter’s reliability and mean time to failure (MTTF) are evaluated using the Markov reliability model, considering all components’ short and open circuit faults. The failure rates of converter components are calculated using the MIL-HDBK-217 methodology. The results obtained from the NSGA-II and HBA are compared to identify the superior algorithm with the best-optimized parameters. Simulation results indicate that the proposed optimization technique enhances the converter’s reliability performance while maintaining its efficiency. Furthermore, this paper presents comprehensive analyses, comparisons, and experimental findings to demonstrate the converter’s fault-tolerant capability. The multi-objective optimization approach offers a practical and effective strategy for optimizing switching power supplies for efficiency and reliability

Incorporating experts’ judgment into machine learning models

Machine learning (ML) models have been quite successful in predicting outcomes in many applications. However, in some cases, domain experts might have a judgment about the expected outcome that might conflict with the prediction of ML models. One main reason for this is that the training data might not be totally representative of the population. In this paper, we present a novel framework that aims at leveraging experts’ judgment to mitigate the conflict. The underlying idea behind our framework is that we first determine, using a generative adversarial network, the degree of representation of an unlabeled data point in the training data. Then, based on such degree, we correct the machine learning model’s prediction by incorporating the experts’ judgment into it, where the higher that aforementioned degree of representation, the less the weight we put on the expert intuition that we add to our corrected output, and vice-versa. We perform multiple numerical experiments on synthetic data as well as two real-world case studies (one from the IT services industry and the other from the financial industry). All results show the effectiveness of our framework; it yields much higher closeness to the experts’ judgment with minimal sacrifice in the prediction accuracy, when compared to multiple baseline methods. We also develop a new evaluation metric that combines prediction accuracy with the closeness to experts’ judgment. Our framework yields statistically significant results when evaluated on that metric.

Can We Accurately Predict Critical Power and W' from a Single Ramp Incremental Exercise Test?

The purpose of this study was to examine the suitability of a single ramp incremental test to predict critical power (CP) and W' . We hypothesized that CP would correspond to the corrected power output (PO) at the respiratory compensation point (RCP) and W' would be calculable from the work done above RCP. One hundred fifty-three healthy young people (26 ± 4 yr, 51.4 ± 7.6 mL·min -1 ·kg -1 ) performed a maximal ramp test (20, 25, or 30 W·min -1 ), followed by three to five constant load trials to determine CP and W' . CP and W' were estimated using a "best individual fit" approach, selecting the mathematical model with the smallest total error. The RCP was identified by means of gas exchange analysis and then translated into its appropriate PO by applying a correction strategy in order to account for the gap in the V̇O 2 /PO relationship between ramp and constant load exercise. We evaluated the agreement between CP and the PO at RCP, and between W' and the total work done above CP ( W'RAMP > CP ) and above RCP ( W'RAMP > RCP ) during the ramp test. The CP was significantly higher than the PO at RCP (Δ = 8 ± 16 W, P < 0.001). W'RAMP > CP was significantly lower than W' (Δ = 1.9 ± 3.3 kJ, P < 0.001), whereas W'RAMP > RCP and W' did not differ from each other (Δ = -0.6 ± 5.8 kJ, P = 0.21). Despite the fact that CP and RCP occurred in close proximity, the estimation of W' from ramp exercise may be problematic given the likelihood of underestimation and considering the large variability. Therefore, we do not recommend the interchangeable use of CP and W' values derived from constant load versus ramp exercise, in particular, when the goal is to obtain accurate estimates or to predict performance capacity.

Bird sound classification based on ECOC-SVM

Birds are an important part of the ecosystem. Knowing the number of existing bird species is important for maintaining ecological balance. The sounds of birds contain rich biological features. The classification of bird species based on their sounds helps to protect the ecological environment. In this study, a fused classification method based on the Error Correction Output Coding (ECOC) and Support Vector Machines (SVM) is investigated to classify 11 species of birds. The Mel Frequency Cepstrum Coefficients (MFCC) of bird sound are extracted as acoustic feature. The ECOC-SVM is compared with Random Forest (RF), Gaussian Mixture Model (GMM) and a multi-layer perceptron neural network (MLP-NN) and a convolutional neural network (CNN) for the bird sound classification. The proposed ECOC-SVM model is validated on the open dataset available from the Macaulay Library at Cornell University. The ECOC-SVM achieves the best classification result and relatively low computational cost among all discussed classifiers on the test set with 100 % accuracy and 52 ms prediction time. The results show that the proposed ECOC-SVM model could well apply on the bird sound classification of multiple species with excellent performance and relatively low computational cost.

Quantitative Analysis of the Effects of Environmental Magnetic Field in Gravity Gradient Sensing

Magnetic field affects the sensing performance of rotating accelerometer gravity gradient instrument (RAGGI) used for airborne mineral exploration and gravity-aided navigation. This article studies the error introduction mechanism of the RAGGI measurement platform caused by the environmental magnetic field. It clarifies the inseparable function relationship between gravity gradient and magnetic gradient based on the magnetic-feedback-accelerometer RAGGI measurement principle. The RAGGI gravity gradient sensing output with environmental magnetic field quantitative analysis model is proposed and established. The experimental verification scheme of moving a ferromagnetic gravitational source for generating the gravity gradient anomaly and magnetic gradient anomaly simultaneously is set up. The experimental results witness that by lowering the magnetic sensitivity coefficient of the four accelerometers from >10 to < 0.1 V/T, the proportion of RAGGI output caused by magnetic gradient anomaly is successfully decreased from >56% to < 1% and verifies the effectiveness of the proposed error correction output model.

A Hybridized Artificial Neural Network for Automated Software Test Oracle

Software testing is the methodology of analyzing the nature of software to test if it works as anticipated so as to boost its reliability and quality. These two characteristics are very critical in the software applications of present times. When testers want to perform scenario evaluations, test oracles are generally employed in the third phase. Upon test case execution and test outcome generation, it is essential to validate the results so as to establish the software behavior’s correctness. By choosing a feasible technique for the test case optimization and prioritization as along with an appropriate assessment of the application, leads to a reduction in the fault detection work with minimal loss of information and would also greatly reduce the cost for clearing up. A hybrid Particle Swarm Optimization (PSO) with Stochastic Diffusion Search (PSO-SDS) based Neural Network, and a hybrid Harmony Search with Stochastic Diffusion Search (HS-SDS) based Neural Network has been proposed in this work. Further to evaluate the performance, it is compared with PSO- SDS based artificial Neural Network (PSO-SDS ANN) and Artificial Neural Network (ANN). The Misclassification of correction output (MCO) of HS-SDS Neural Network is 6.37 for 5 iterations and is well suited for automated testing.

Research on Bidirectional Isolated Charging System Based on Resonant Converter

This paper proposes a two-stage bidirectional isolated charging system, which can realize the bidirectional flow of electric energy, not only making the electric energy flow from the grid side to the battery side but also converting the battery’s energy into single-phase alternating current to supply other electric equipment. The charging system also has the function of power factor correction and wide-range voltage output. The front stage of the charging system is a bidirectional totem pole structure power factor correction converter with a voltage and current double closed-loop control strategy to ensure the stability of the DC bus voltage, and the rear stage is a bidirectional CLLLC resonant converter, which adopts a high-frequency soft start strategy to reduce the inrush current during start-up and ensure the safe operation of the converter. Moreover, the frequency control strategy is used to make it have a wide range of DC output characteristics. In this paper, the principle analysis and parameter calculation of the charging system is carried out, the simulation platform and hardware circuit design are built, and a test prototype is piloted to verify the bidirectional operating characteristics of the charging system. The simulation and experimental results demonstrate the correctness of the theoretical analysis and topology design.

Repairing geometric errors in 3D urban models with kinetic data structures

3D urban models created either interactively by human operators or automatically with reconstruction algorithms often contain geometric and semantic errors. Correcting them in an automated manner is an important scientific challenge. Prior work, which traditionally relies on local analysis and heuristic-based geometric operations on mesh data structures, is typically tailored-made for specific 3D formats and urban objects. We propose a more general method to process different types of urban models without tedious parameter tuning. The key idea lies on the construction of a kinetic data structure that decomposes the 3D space into polyhedra by extending the facets of the imperfect input model. Such a data structure allows us to re-build all the relations between the facets in an efficient and robust manner. Once built, the cells of the polyhedral partition are regrouped by semantic classes to reconstruct the corrected output model. We demonstrate the robustness and efficiency of our algorithm on a variety of real-world defect-laden models and show its competitiveness with respect to traditional mesh repairing techniques from both Building Information Modeling (BIM) and Geographic Information Systems (GIS) data.

Diabetic Retinopathy Detection Using Genetic Algorithm-Based CNN Features and Error Correction Output Code SVM Framework Classification Model

Diabetic retinopathy (DR) is a type of eye disease that may be caused in individuals suffering from diabetes which results in vision loss. DR identification and routine diagnosis is a challenging task and may need several screenings. Early identification of DR has the potential to prevent or delay vision loss. For real-time applications, an automated DR identification approach is required to assist and reduce possible human mistakes. In this research work, we propose a deep neural network and genetic algorithm-based feature selection approach. Five advanced convolutional neural network architectures are used to extract features from the fundus images, i.e., AlexNet, NASNet-Large, VGG-19, Inception V3, and ShuffleNet, followed by genetic algorithm for feature selection and ranking features into high rank (optimal) and lower rank (unsatisfactory). The nonoptimal feature attributes from the training and validation feature vectors are then dropped. Support vector machine- (SVM-) based classification model is used to develop diabetic retinopathy recognition model. The model performance is evaluated using accuracy, precision, recall, and F1 score. The proposed model is tested on three different datasets: the Kaggle dataset, a self-generated custom dataset, and an enhanced custom dataset with 97.9%, 94.76%, and 96.4% accuracy, respectively. In the enhanced custom dataset, data augmentation has been performed due to the smaller size of the dataset and to eliminate the noise in fundus images.

Biometric Information Recognition Using Artificial Intelligence Algorithms: A Performance Comparison

Addressing crime detection, cyber security and multi-modal gaze estimation in biometric information recognition is challenging. Thus, trained artificial intelligence (AI) algorithms such as Support vector machine (SVM) and adaptive neuro-fuzzy inference system (ANFIS) have been proposed to recognize distinct and discriminant features of biometric information (intrinsic hand features and demographic cues) with good classification accuracy. Unfortunately, due to nonlinearity in distinct and discriminant features of biometric information, accuracy of SVM and ANFIS is reduced. As a result, optimized AI algorithms ((ANFIS) with subtractive clustering (ANFIS-SC) and SVM with error correction output code (SVM-ECOC)) have shown to be effective for biometric information recognition. In this paper, we compare the performance of the ANFIS-SC and SVM-ECOC algorithms in their effectiveness at learning essential characteristics of intrinsic hand features and demographic cues based on Pearson correlation coefficient (PCC) feature selection. Furthermore, the accuracy of these algorithms are presented, and their recognition performances are evaluated by root mean squared error (RMSE), mean absolute percentage error (MAPE), scatter index (SI), mean absolute deviation (MAD), coefficient of determination (R2), Akaike’s Information Criterion (AICc) and Nash-Sutcliffe model efficiency index (NSE). Evaluation results show that both SVM-ECOC and ANFIS-SC algorithms are suitable for accurately recognizing soft biometric information on basis of intrinsic hand measurements and demographic cues. Moreover, comparison results demonstrated that ANFIS-SC algorithms can provide better recognition accuracy, with RMSE, AICc, MAPE, R2 and NSE values of ≤ 3.85, 2.39E+02, 0.18%, ≥ 0.99 and ≥ 99, respectively.

Investigation of field output factors using IAEA-AAPM TRS-483 code of practice recommendations and Monte Carlo simulation for 6 MV photon beams

Abstract Introduction: This study aims to experimentally determine field output factors using the methodologies suggested by the IAEA-AAPM TRS-483 for small field dosimetry and compare with the calculation from Monte Carlo (MC) simulation. Methods: The IBA-CC01, Sun Nuclear EDGE and IBA-SFD detectors were employed to determine the uncorrected and the corrected field output factors for 6 MV photon beams. Measurements were performed at 100 cm source to axis distance, 10 cm depth in water, and the field sizes ranged from 1 × 1 to 10 × 10 cm2. The use of field output correction factors proposed by the TRS-483 was utilised to determine field output factors. The measured field output factors were compared to that calculated using the egs_chamber user code. Results: The decrease in the percentage standard deviation of the measured three detectors was observed after applying the field output correction factors. Measured field output factors using CC01 and EDGE detectors agreed with MC values within 3% for field sizes down to 1 × 1 cm2, except the SFD detector. Conclusions: The corrected field output factors agree with the calculation from MC, except the SFD detector. CC01 and EDGE are suitable for determining field output factors, while the SFD may need more implementation of the intermediate field method.

Non-uniform electric field in cantilevered piezoelectric energy harvesters: An improved distributed parameter electromechanical model

In the existing electromechanical model for cantilevered piezoelectric energy harvesters, the electric field in the piezoelectric layer is generally considered as a uniform field. In the present study, it is found and described in detail that the oversimplified uniform electric field assumption could lead to critical issues regarding the fundamental of piezoelectricity. Further, inspired by the electric field distribution obtained using the exact solutions for plane problems, a non-uniform electric field is proposed. Accordingly, an improved distributed parameter electromechanical model for cantilevered piezoelectric energy harvesters including both unimorph and bimorph configurations is established, in which the bending stiffness term involved in the electromechanical governing equations is explicitly corrected, and the aforementioned critical issues are perfectly addressed. The present proposed non-uniform electric field is then validated by comparing the static analysis results of the present model with finite element results and the results obtained using exact solutions for plane problems. Dynamic analysis of the present model is also conducted, and the electromechanical characteristic corrections for the traditional distributed parameter model adopting uniform electric field assumption is carried out including bending stiffness correction, resonant frequency correction, voltage output correction and power output correction.

Neural OCR Post-Hoc Correction of Historical Corpora

Abstract Optical character recognition (OCR) is crucial for a deeper access to historical collections. OCR needs to account for orthographic variations, typefaces, or language evolution (i.e., new letters, word spellings), as the main source of character, word, or word segmentation transcription errors. For digital corpora of historical prints, the errors are further exacerbated due to low scan quality and lack of language standardization. For the task of OCR post-hoc correction, we propose a neural approach based on a combination of recurrent (RNN) and deep convolutional network (ConvNet) to correct OCR transcription errors. At character level we flexibly capture errors, and decode the corrected output based on a novel attention mechanism. Accounting for the input and output similarity, we propose a new loss function that rewards the model’s correcting behavior. Evaluation on a historical book corpus in German language shows that our models are robust in capturing diverse OCR transcription errors and reduce the word error rate of 32.3% by more than 89%.