Search

Aim: To determine the hemodynamic response during insertion of laryngeal mask airway versus conventional intubation. Study design: Randomized controlled trial. Place and duration of study: Department of Anaesthesia, Jinnah Postgraduate Medical Centre, Karachi from 9th June 2016 to 10th December 2016. Methodology: One hundred and fifty-eight patients were enrolled, and they were divided in two groups; Group A (laryngeal mask airway) and patients falling in group B (conventional method). Baseline haemodynamic parameters were noted, and all patients were induced with propofol 2mg.kg 1. In group B, succinylcholine 1.5mg.kg-1 was used to facilitate intubation. After induction appropriate size endotracheal tube or laryngeal mask airway were inserted for airway control. For first five minutes after intervention, analgesics and any other stimulation were avoided, in order to prevent any haemodynamic alteration. All airway interventions were done by anaesthesiologist who had more than 5 years post fellowship experience. Mean arterial pressures were recorded. Initial haemodynamic parameters were measured when the patient enter the operating room and then second reading taken just after induction of anaesthesia, third reading recorded one minute and fourth reading 5 minutes after the intervention (i.e., after passing either endotracheal tube or laryngeal mask airway). Results: The average mean arterial pressure during process of intubation of patients in Group laryngeal mask airway group was 105.21±5.90 while in conventional group the average mean arterial pressure was 102.21±4.29 with P-value=0.001. Conclusion: Intubation through intubating laryngeal mask airway is accompanied by minimal cardiovascular responses than those associated with direct laryngoscopic tracheal intubation, so it can be used for patients in whom a marked pressor response would be deleterious. Keywords: Intubating laryngeal mask airway, Conventional laryngoscopy, Hemodynamic responses, Airway morbidity

Background: Fentanyl-induced cough is a common difficulty encountered at induction of anesthesia. Various interventions both pharmacological and non-pharmacological have been used to prevent this side effect including dexamethasone and propofol. Aim: To compare the effectiveness of dexamethasone and propofol to prevent fentanyl-induced cough at the induction of general anesthesia. Study design: Randomized controlled trial. Place and duration of study: Department of Anaesthesiology, Surgical ICU & Pain Management, Dow University of Health Sciences and Civil Hospital Karachi from 16th September 2011 to 15th March, 2012. Methodology: One hundred patients who underwent elective surgical procedure were selected. Patients were randomized in two groups of 50 patients each; Group D was given intravenous dexamethasone; whereas patients of group P received intravenous propofol as the premedication before induction. The main outcome measure was effectiveness of both drugs to prevent fentanyl-induced cough. Results: Majority of patients (40%) were between 20-30 years of age group with mean age was 35.80±10.14 years. Males were more than females. Intravenous dexamethasone was significantly effective (90%) than intravenous propofol (70%) [p=0.012]. Conclusion: Intravenous dexamethasone is effective in reducing fentanyl-induced cough in comparison to propofol. Keywords: Dexamethasone, Propofol, Fentanyl-induced cough (FIC)

Background: There are currently no treatment algorithms specifically for early peri-prosthetic joint infection (PJI) after hemiarthroplasty for hip fracture. Commonly, debridement, antibiotics, and implant retention (DAIR) is attempted as first-line management, despite lack of evidence supporting this strategy in this patient group. The purpose of this study was to evaluate outcomes of DAIR for early PJI after hemiarthroplasty for hip fracture in our unit. Methods: The departmental database from December 2008 to January 2019 was searched to identify all patients in our unit who were treated for early PJI after hemiarthroplasty for hip fracture. Data for included patients were collected from electronic healthcare records and analyzed. Primary outcome measure was treatment success, defined as patient survival to discharge, with eradication of infection and implant retention. Results: Twenty-six patients were identified and included in the study. Mean age was 84.7 years. All except one patient were American Society of Anesthesiologists (ASA) class 3 or 4. All patients were McPherson host grade B or C. Twenty-three of 26 patients underwent DAIR and three of 26 proceeded directly to excision arthroplasty. Debridement, antibiotics, and implant retention was successful in three of 23 patients (13%) after a single procedure, with success in two additional patients after a second procedure, giving overall success rate of five of 23 patients (22%). Conclusions: Debridement, antibiotics, and implant retention has a high failure rate in treating early PJI after hemiarthroplasty for hip fracture. These patients are generally elderly and frail with multiple host and wound compromising factors. Debridement, antibiotics, and implant retention should not be recommended as first-line management for the majority of these patients, for whom getting it right the first time is of vital importance to avoid consequences associated with failed surgical procedures. Further multicenter studies that also explore alternate treatment strategies are required to devise an algorithm specifically for hip fracture patients, to aid decisions on treatment and improve outcomes.

Electronic Health Records (EHR) containing longitudinal information about millions of patient lives are increasingly being utilized by organizations across the healthcare spectrum. Studies on EHR data have enabled real world applications like understanding of disease progression, outcomes analysis, and comparative effectiveness research. However, often every study is independently commissioned, data is gathered by surveys or specifically purchased per study by a long and often painful process. This is followed by an arduous repetitive cycle of analysis, model building, and generation of insights. This process can take anywhere between 1 - 3 years. In this paper, we present a robust end-to-end machine learning based SaaS system to perform analysis on a very large EHR dataset. The framework consists of a proprietary EHR datamart spanning ~55 million patient lives in USA and over ~20 billion data points. To the best of our knowledge, this framework is the largest in the industry to analyze medical records at this scale, with such efficacy and ease. We developed an end-to-end ML framework with carefully chosen components to support EHR analysis at scale and suitable for further downstream clinical analysis. Specifically, it consists of a ridge regularized Survival Support Vector Machine (SSVM) with a clinical kernel, coupled with Chi-square distance-based feature selection, to uncover relevant risk factors by exploiting the weak correlations in EHR. Our results on multiple real use cases indicate that the framework identifies relevant factors effectively without expert supervision. The framework is stable, generalizable over outcomes, and also found to contribute to better out-of-bound prediction over known expert features. Importantly, the ML methodologies used are interpretable which is critical for acceptance of our system in the targeted user base. With the system being operational, all of these studies were completed within a time frame of 3-4 weeks compared to the industry standard 12-36 months. As such our system can accelerate analysis and discovery, result in better ROI due to reduced investments as well as quicker turn around of studies.

Approximately 350 million patients of hepatitis B and 170 million patients of Hepatitis C are present worldwide according to WHO. Many risk factors are involved in the transmission of theses deadly viral infections but blood transfusion in Beta thalassemic patients is working with two faces, one as remedy and the other is key risk factor in the spread of silent killers. Thalassemia patients registered in Combine Military Hospital (CMH) Rawalakot and Sheikh Khalifa Bin Zayed Al-Nahyan Hospital, Muzaffarabad Azad Jammu and Kashmir Pakistan were studied for the viral hepatitis B and C prevalence. A total of 303 (including 164 males and 139 females) individuals, aged between 1 and 12 years were studied. All the understudy participants were interviewed through questionnaire method. After taking written consent from each participant or guardian, 5 ml of blood was collected from each participant and brought to the working laboratory for HBV and HCV screening through ICT kit method. All ICT positive samples were further confirmed through ELISA. Individuals 25(8.2%) were found positive for both hepatitis B surface Antigen (HBsAg) and Anti hepatitis C antibody (Anti-HCV antibody) after initial screening with no coinfection of both diseases. Out of 25 total infected individuals, 05(1.6%) were found HBsAg positive and 20(6.6%) were found anti-HCV positive. All the ICT positive individuals were further confirmed by quantitative Enzyme Linked Immunosorbent Assay (ELISA) and 23(7.6%) individuals were confirmed for both hepatitis B and C including 05(1.6%) HBsAg positive as well as 18(5.9%) anti-HCV antibody positive individuals. We can conclude that 8.2% prevalence of hepatitis B and C among thalassemic patients is an alarming health concern which directly indicates to pay attention for ensuring 100% safe blood transfusion.

Despite growing concerns over pathological internet usage, studies based on validated psychometric instruments are still lacking in Pakistan. This study aimed to examine the psychometric properties of the Internet Addiction Test (IAT) in a sample of Pakistani students. A total of 522 students of medicine and dentistry completed the questionnaire, which consisted of four sections: (a) demographics, (b) number of hours spent on the Internet per day, (c) English version of the IAT, and (d) the Defense Style Questionnaire-40. Maximum likelihood analysis and principal axis factoring were used to validate the factor structure of the IAT. Convergent and criterion validity were assessed by correlating IAT scores with number of hours spent online and defense styles. Exploratory and confirmatory factor analysis reflected the goodness of fit of a unidimensional structure of the IAT, with a high alpha coefficient. The IAT had good face and convergent validity and no floor and ceiling effects, and was judged easy to read by participants.

This paper presents a new method to recognize human facial expression by feeding hybrid features to Self-organizing maps (SOM). Facial expression recognition is still a challenging problem that can be seen in many real applications such as security systems, behavioral science and clinical practices. In this work, we present a new way to analyze, represent and recognize human facial expressions from a video sequence. The proposed facial expression recognition framework comprises of following components: face detection, facial feature extraction and classification. Firstly, faces are detected based on skin color after removing background and noise effects from raw video sequences. Then each face image is aligned using vertex mask generation and the 1D transformation features are extracted to utilize the local information for each facial image. After reducing the dimension of features and doing independent component analysis, the new features are trained and tested using SOM. The experimental evaluation demonstrates the proposed approach over two public datasets such as Cohn-Kanade and AT&T datasets of facial expression videos achieves superior recognition performance of 96.81% and 96.55% over state of the art methods.

Introduction: Despite growing concerns over pathological use of the Internet, studies based on validated psychometric instruments are still lacking in Pakistan. The present study aimed to examine the psychometric properties of Young’s Internet Addiction Test (IAT) in a sample of the Pakistani population. We examined the validity, internal consistency, readability and floor and ceiling effects of IAT scores.Methods: This cross-sectional study was conducted at CMH Lahore Medical College and Institute of Dentistry, Lahore, Pakistan from 1 March 2015 to 30 May 2015. A total of 522 medical and dental students completed the questionnaire, which consisted of three sections: (a) demographics and percentage grades in annual examinations, (b) a categorical question to record the estimated number of hours spent on the Internet per day, and (c) the English version of the IAT. All data were analyzed in SPSS v. 20. Principal axis factor analysis was used to validate the factor structure of the IAT in our study sample. An alpha coefficient > .7 was sought in the reliability analysis. Histograms and the values of skewness and kurtosis were analyzed for floor and ceiling effects. In addition, readability of the IAT was assessed as the Flesch Reading Ease score and Flesch-Kincaid Grade level function. Results: A total of 522 medical and dental students participated in the survey. Most respondents were female medical students enrolled in preclinical years of their degree program. Median age (min-max) of the respondents was 20 years (17-25 years). A single-factor model for IAT score explained 33.71% of the variance, with a high alpha coefficient of .893. In addition, the IAT had good face and convergent validity and no floor and ceiling effects, and was judged easy to read by participants. Conclusion: The English version of the IAT showed good psychometric properties in a sample of Pakistani university students. A single-factor model for assessing internet addiction showed good reliability and was found suitable with our study sample.

Despite several decades of research in document analysis, recognition of unconstrained handwritten documents is still considered a challenging task. Previous research in this area has shown that word recognizers perform adequately on constrained handwritten documents which typically use a restricted vocabulary (lexicon). But in the case of unconstrained handwritten documents, state-of-the-art word recognition accuracy is still below the acceptable limits. The objective of this research is to improve word recognition accuracy on unconstrained handwritten documents by applying a post-processing or OCR correction technique to the word recognition output. In this paper, we present two different methods for this purpose. First, we describe a lexicon reduction-based method by topic categorization of handwritten documents which is used to generate smaller topic-specific lexicons for improving the recognition accuracy. Second, we describe a method which uses topic-specific language models and a maximum-entropy based topic categorization model to refine the recognition output. We present the relative merits of each of these methods and report results on the publicly available IAM database.

Securing biometrics databases from being compromised is an important research challenge that must be overcome in order to support widespread use of biometrics based authentication. In this paper we present a novel method for securing fingerprints by hashing the fingerprint minutia and performing matching in the hash space. Our approach uses a family of symmetric hash functions and does not depend on the location of the (usually unstable) singular points (core and delta) as is the case with other methods described in the literature. It also does not assume a pre-alignment between the test and the stored fingerprint templates. We argue that these assumptions, which are often made, are unrealistic given that fingerprints are very often only partially captured by the commercially available sensors. The Equal Error Rate (EER) achieved by our system is 3%. We also present the performance analysis of a hybrid system that has an EER of 1.96% which reflects almost no drop in performance when compared to straight matching with no security enhancements. The hybrid system involves matching using our secure algorithm but the final scoring reverts to that used by a straight matching system.

ObjectiveThe ability to predict patient readmission risk is extremely valuable for hospitals, especially under the Hospital Readmission Reduction Program of the Center for Medicare and Medicaid Services which went into effect starting October 1, 2012. There is a plethora of work in the literature that deals with developing readmission risk prediction models, but most of them do not have sufficient prediction accuracy to be deployed in a clinical setting, partly because different hospitals may have different characteristics in their patient populations. Methods and materialsWe propose a generic framework for institution-specific readmission risk prediction, which takes patient data from a single institution and produces a statistical risk prediction model optimized for that particular institution and, optionally, for a specific condition. This provides great flexibility in model building, and is also able to provide institution-specific insights in its readmitted patient population. We have experimented with classification methods such as support vector machines, and prognosis methods such as the Cox regression. We compared our methods with industry-standard methods such as the LACE model, and showed the proposed framework is not only more flexible but also more effective. ResultsWe applied our framework to patient data from three hospitals, and obtained some initial results for heart failure (HF), acute myocardial infarction (AMI), pneumonia (PN) patients as well as patients with all conditions. On Hospital 2, the LACE model yielded AUC 0.57, 0.56, 0.53 and 0.55 for AMI, HF, PN and All Cause readmission prediction, respectively, while the proposed model yielded 0.66, 0.65, 0.63, 0.74 for the corresponding conditions, all significantly better than the LACE counterpart. The proposed models that leverage all features at discharge time is more accurate than the models that only leverage features at admission time (0.66 vs. 0.61 for AMI, 0.65 vs. 0.61 for HF, 0.63 vs. 0.56 for PN, 0.74 vs. 0.60 for All Cause). Furthermore, the proposed admission-time models already outperform the performance of LACE, which is a discharge-time model (0.61 vs. 0.57 for AMI, 0.61 vs. 0.56 for HF, 0.56 vs. 0.53 for PN, 0.60 vs. 0.55 for All Cause). Similar conclusions can be drawn from other hospitals as well. The same performance comparison also holds for precision and recall at top-decile predictions. Most of the performance improvements are statistically significant. ConclusionsThe institution-specific readmission risk prediction framework is more flexible and more effective than the one-size-fit-all models like the LACE, sometimes twice and three-time more effective. The admission-time models are able to give early warning signs compared to the discharge-time models, and may be able to help hospital staff intervene early while the patient is still in the hospital.

The translation of non-invasive Brain-Computer Interfaces (BCI) from the only possible communication pathway for paralyzed patients into more widespread applications is limited by: the lack of effective, user-friendly and robust paradigms, low information transfer rates (ITR), and lack of suitable electroencephalography (EEG) recording platforms. This study addresses the last point by exploring the extent to which the recently introduced EarEEG technology, which provides an unobtrusive, discreet and user-friendly way of recording EEG, can be used in BCI applications. This was approached by comparing conventional on-scalp EEG and EarEEG recordings in a well-established visual P300 BCI setup. The two recording methods were compared qualitatively by comparing Event-Related Potential (ERP) waveforms, and quantitatively in terms of P300 signal-to-noise ratios (SNR) and ITRs of the BCI paradigm. The study showed similar ERP waveforms and on par P300 SNRs from both on-scalp and ear electrodes, and only a 6.5% decrease in single channel ITR for ear electrodes compared to on-scalp electrodes. This demonstrates that the EarEEG platform is a feasible technology for P300-based BCIs.

With a growing focus on renewable energy, interest in design of wind turbines has also been expanding. In today’s market, the horizontal axis (windmill) turbine is the most common type in use; but, vertical axis (Darrieus) turbines have certain advantages. Darrieus turbines, which are liftdriven, have a higher power potential than the horizontal, or drag-driven turbines. The main flaw with their design is their inability to self-start. Darrieus turbines require an external energy source to bring the device to a minimum rotational speed. This paper presents design, construction and testing of a vertical axias (Darrieus) wind turbine with 3 blades, starting solely from the low energy of the wind. A separate drag device (Savonius type turbine) on the top of an existing Darrieus turbine was mounted to make the turbine self start at low wind speed. The cut-in speed of the turbine was 3 m/s, cut-off speed was 20 m/s and power obtained was 50 Watts at wind velocity of 6 m/s. The testing with primiarly available permanent magnetic generator sponsored by industry resulted in 35 Watts at 9m/s. doi: 10.5829/idosi.ijee.2015.06.01.02

This workshop report is an overview of the Predictive Models in Personalized Medicine workshop held on Dec. 11, 2010 at 2010 Neural Information Processing Systems (NIPS) Conference in Whistler, Canada. The workshop included 3 keynote talks and 6 oral and 5 poster presentations on peer reviewed submissions. The workshop also featured a panel discussion on the growing trends of the field. The workshop was very well received and attended. Based on the response and multiple requests received, it is planned to organize the workshop again at NIPS 2011.

The ability to predict patient readmission risk is extremely valuable for hospitals, especially under the Hospital Readmission Reduction Program (HRRP) of the Center for Medicare and Medicaid Services (CMS) which went into effect starting October 1, 2012. There is a plethora of work in the literature that deals with developing readmission risk prediction models, but most of them do not have sufficient prediction accuracy to be deployed in a clinical setting, partly because different hospitals may have different characteristics in their patient populations. In this work we experimented with a generic framework for institution-specific readmission risk prediction, which takes patient data from a single institution and produces a statistical risk prediction model optimized for that particular institution and optionally condition specific. This provides great flexibility in model building, and is also able to provide institution-specific insights in its readmitted patient population. We showcase some initial results at three institutions for Heart Failure (HF), Acute Myocardial Infarction (AMI) and Pneumonia (PN) patients. The developed models yield better prediction accuracy than the ones present in the literature.

Feature extraction is one of the most important modules for Facial Expression Recognition (FER) systems, which deals with getting the distinguishable features each expression and quantizing it as a discrete symbol. In this paper, we have proposed the novel robust feature extraction technique for the FER systems called Stepwise Linear Discriminant Analysis (SWLDA). This technique focuses on the selection of localized features from the facial expression images and discriminate their classes on the basis of regression values i.e. partial F-test. The proposed technique is then compared with conventional techniques such as LDA in combination with ICA. The results shows that SWLDA better than conventional techniques in terms of robustness in suitable feature selection and classification.

Minutiae based fingerprint matching algorithms are wildly used in fingerprint identification and verification applications. However, they may suffer from spurious matches because they do not use the rich local image information. In this paper, we extend minutiae based methods to incorporate such local image information. Our method uses local mutual information, a proven similarity measure in various applications, to improve the matching rate. The overall minutiae distribution pattern between two fingerprints is represented by the initial minutiae matching result, while the mutual information measures the similarity between neighborhoods of matched minutiae, thus enhancing the final matching decision. FVC2002 DB1 and DB3 databases are used to test the proposed approach. Experimental result shows the improvement when combining minutiae matching scores with mutual information scores.

Despite several decades of research in handwriting recognition, the goal of having computers access handwritten information from unconstrained document images is still elusive. Current handwriting recognition systems are only capable of recognizing words that are present in a restricted lexicon typically comprised of 10 to 1000 words. As the size of the lexicon grows, the recognition accuracy falls sharply and is reported to be around 30% for a10K word lexicon. The objective of this research is to raise the accuracy levels on unconstrained handwritten documents by reducing the size of lexicons. We present an innovative method of lexicon reduction by topic categorization of handwritten documents. After categorization of a document into a topic e.g. sports, science etc. we use smaller lexicons that include only words with high mutual information with that topic and hence increase performance of recognizers. In this paper we present different techniques and report results on a publicly available dataset.

Despite several decades of research in document analysis, recognition of unconstrained handwritten documents is still considered a challenging task. Previous research in this area has shown that word recognizers produce reasonably clean output when used with a restricted lexicon. But in absence of such a restricted lexicon, the output of an unconstrained handwritten word recognizer is noisy. The objective of this research is to process noisy recognizer output and eliminate spurious recognition choices using a topic based language model. We construct a topic based language model for every document using a training data which is manually categorized. A topic categorization sub-system based on Maximum Entropy model is also trained which is used to generate the topic distribution of a test document. A given test word image is processed by the recognizer and its word recognition likelihood is refined by incorporating topic distribution of the document and topic based language model probability. The proposed method is evaluated on a publicly available IAM dataset and experimental results show significant improvement in the word recognition accuracy from 32% to 40% over a test set consisting of 4033 word images extracted from 70 handwritten document images.

Biometric identification has numerous advantages over conventional ID and password systems; however, the lack of anonymity and revocability of biometric templates is of concern. Several methods have been proposed to address these problems. Many of the approaches require a precise registration before matching in the anonymous domain. We introduce binary string representations of fingerprints that obviates the need for registration and can be directly matched. We describe several techniques for creating anonymous and revocable representations using these binary string representations. The match performance of these representations is evaluated using a large database of fingerprint images. We prove that given an anonymous representation, it is computationally infeasible to invert it to the original fingerprint, thereby preserving privacy. To the best of our knowledge, this is the first linear, anonymous and revocable fingerprint representation that is implicitly registered.

Automatic language identification is an important step prior to optical character recognition (OCR). In this paper we present a system to discriminate between Arabic and Persian in historical Afghan manuscripts. The classification is performed at a sub-sentence level. We propose a feature extraction algorithm for a sub-sentence based on Gabor filters followed by classification using a support vector machine (SVM). An overall precision of 96.72% and 94.90% is obtained for Persian and Arabic respectively.

In this paper we present a system for classification of machine printed and handwritten text in mixed documents. The classification is performed at the word level. We propose a feature extraction algorithm for each word image based on Gabor filters followed by classification using an expectation maximization (EM) based probabilistic neural network that reduces overfitting of training data. An overall precision of 94.62% was obtained for the Arabic script using the modified neural network. The accuracies obtained using a simple backpropagation neural network and an SVM were 83.33% and 90.26% respectively

In order to improve the readability and the automatic recognition of handwritten document images, preprocessing steps are imperative. These steps in addition to conventional steps of noise removal and filtering include text normalization such as baseline correction, slant normalization and skew correction. These steps make the feature extraction process more reliable and effective. Recently Arabic handwriting recognition has received some attention from the research community. Due to the unique nature of the script, the conventional methods do not prove to be effective. In our work, we describe an orientation independent technique for baseline detection of Arabic words. In addition to that we describe, in the rest of the paper, our techniques for slant normalization, slope correction, line and word separation in handwritten Arabic documents. We show how the baseline can be exploited for slope and skew correction before proceeding with the steps of line and word separation.

The possibility that a biometric database is compromised is one of the main concerns in implementing biometric identification systems. The compromise of a biometric renders it permanently useless. In this paper we present a method of hashing fingerprint minutia information and performing fingerprint identification in a new space. Only hashed data is transmitted and stored in the server database, and it is not possible to restore fingerprint minutia locations using hashed data. We also present a performance analysis of the proposed algorithm.

We present methodologies for three important tasks that will eventually enable digital access of multilingual Indian document images. First, we describe several document image analysis techniques necessary to prepare Devanagari document images for OCR. The second task is OCR for machine printed Devanagari words without the help of a lexicon. We describe the OCR methodology and show how it is being extended to other Indian languages. Finally, we describe a versatile platform that facilitates automatic segmentation of document images in multiple Indian languages and an interface to capture the ground truth corresponding to the text. We use transliterated English text and virtual keyboards in a range of Indian languages for this purpose. The multilingual data entry capabilities of the tool and its underlying UNICODE data representation within a structured XML document also allow users to annotate passages of text in one language in other languages using a markup scheme to switch between scripts. Text and annotations are rendered in the appropriate scripts as the text is being annotated, thus providing users prompt and natural feedback. The XML back-end allows meta-data to be recorded describing the annotated document.

To encourage the delivery of high quality care with an additional emphasis on the transparency of hospital results and reporting, the Centers for Medicare & Medicaid Services (CMS) and the Joint Commission require each hospital to submit certain inpatient core measures every quarter. In this research, we developed a system that can achieve this automatically with high accuracy for any arbitrary clinical question, in a manner that is easily user configurable/teachable.

In Codice Ratio is a research project to study tools and techniques for analyzing the contents of historical documents conserved in the Vatican Secret Archives (VSA). In this paper, we present our efforts to develop a system to support the transcription of medieval manuscripts. The goal is to provide paleographers with a tool to reduce their efforts in transcribing large volumes, as those stored in the VSA, producing good transcriptions for significant portions of the manuscripts. We propose an original approach based on character segmentation. Our solution is able to deal with the dirty segmentation that inevitably occurs in handwritten documents. We use a convolutional neural network to recognize characters and language models to compose word transcriptions. Our approach requires minimal training efforts, making the transcription process more scalable as the production of training sets requires a few pages and can be easily crowdsourced. We have conducted experiments on manuscripts from the Vatican Registers, an unreleased corpus containing the correspondence of the popes. With training data produced by 120 high school students, our system has been able to produce good transcriptions that can be used by paleographers as a solid basis to speedup the transcription process at a large scale., Donatella Firmani, Marco Maiorino, Paolo Merialdo, and Elena Nieddu. 2018. Towards Knowledge Discovery from the Vatican Secret Archives. In Codice Ratio - Episode 1: Machine Transcription of the Manuscripts. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD '18). ACM, New York, NY, USA, 263-272