scholarly journals Incremental Learning for Dermatological Imaging Modality Classification

2021 ◽  
Vol 7 (9) ◽  
pp. 180
Author(s):  
Ana C. Morgado ◽  
Catarina Andrade ◽  
Luís F. Teixeira ◽  
Maria João M. Vasconcelos
Keyword(s):  
Learning Strategies ◽  
Incremental Learning ◽  
Imaging Modality ◽  
Computational Cost ◽  
National Health System ◽  
Considerable Effort ◽  
Experience Replay ◽  
Global Accuracy ◽  
Image Modality

With the increasing adoption of teledermatology, there is a need to improve the automatic organization of medical records, being dermatological image modality a key filter in this process. Although there has been considerable effort in the classification of medical imaging modalities, this has not been in the field of dermatology. Moreover, as various devices are used in teledermatological consultations, image acquisition conditions may differ. In this work, two models (VGG-16 and MobileNetV2) were used to classify dermatological images from the Portuguese National Health System according to their modality. Afterwards, four incremental learning strategies were applied to these models, namely naive, elastic weight consolidation, averaged gradient episodic memory, and experience replay, enabling their adaptation to new conditions while preserving previously acquired knowledge. The evaluation considered catastrophic forgetting, accuracy, and computational cost. The MobileNetV2 trained with the experience replay strategy, with 500 images in memory, achieved a global accuracy of 86.04% with only 0.0344 of forgetting, which is 6.98% less than the second-best strategy. Regarding efficiency, this strategy took 56 s per epoch longer than the baseline and required, on average, 4554 megabytes of RAM during training. Promising results were achieved, proving the effectiveness of the proposed approach.

Incremental learning strategies for credit cards fraud detection: Extended abstract

2020 ◽  
Author(s):  
Bertrand Lebichot ◽  
Gian Marco Paldino ◽  
Gianluca Bontempi ◽  
Wissam Siblini ◽  
Liyun He-Guelton ◽  
...  
Keyword(s):  
Learning Strategies ◽  
Incremental Learning ◽  
Credit Cards ◽  
Fraud Detection

scholarly journals A CAUSAL HIERARCHICAL MARKOV FRAMEWORK FOR THE CLASSIFICATION OF MULTIRESOLUTION AND MULTISENSOR REMOTE SENSING IMAGES

2020 ◽  
Vol V-3-2020 ◽  
pp. 269-277
Author(s):  
A. Montaldo ◽  
L. Fronda ◽  
I. Hedhli ◽  
G. Moser ◽  
S. B. Serpico ◽  
...  
Keyword(s):  
High Resolution ◽  
Markov Chain Model ◽  
Computational Cost ◽  
Order Relation ◽  
Boosted Regression Trees ◽  
Chain Model ◽  
Remotely Sensed Data ◽  
Space Filling Curve ◽  
Very High

Abstract. In this paper, a multiscale Markov framework is proposed in order to address the problem of the classification of multiresolution and multisensor remotely sensed data. The proposed framework makes use of a quadtree to model the interactions across different spatial resolutions and a Markov model with respect to a generic total order relation to deal with contextual information at each scale in order to favor applicability to very high resolution imagery. The methodological properties of the proposed hierarchical framework are investigated. Firstly, we prove the causality of the overall proposed model, a particularly advantageous property in terms of computational cost of the inference. Secondly, we prove the expression of the marginal posterior mode criterion for inference on the proposed framework. Within this framework, a specific algorithm is formulated by defining, within each layer of the quadtree, a Markov chain model with respect to a pixel scan that combines both a zig-zag trajectory and a Hilbert space-filling curve. Data collected by distinct sensors at the same spatial resolution are fused through gradient boosted regression trees. The developed algorithm was experimentally validated with two very high resolution datasets including multispectral, panchromatic and radar satellite images. The experimental results confirm the effectiveness of the proposed algorithm as compared to previous techniques based on alternate approaches to multiresolution fusion.

scholarly journals Watching a Small Portion could be as Good as Watching All: Towards Efficient Video Classification

2018 ◽  
Author(s):  
Hehe Fan ◽  
Zhongwen Xu ◽  
Linchao Zhu ◽  
Chenggang Yan ◽  
Jianjun Ge ◽  
...  
Keyword(s):  
Large Scale ◽  
Sampling Rate ◽  
Computational Cost ◽  
Confidence Score ◽  
Video Content ◽  
Video Classification ◽  
Video Frames ◽  
Similar Accuracy ◽  
Efficient Video

We aim to significantly reduce the computational cost for classification of temporally untrimmed videos while retaining similar accuracy. Existing video classification methods sample frames with a predefined frequency over entire video. Differently, we propose an end-to-end deep reinforcement approach which enables an agent to classify videos by watching a very small portion of frames like what we do. We make two main contributions. First, information is not equally distributed in video frames along time. An agent needs to watch more carefully when a clip is informative and skip the frames if they are redundant or irrelevant. The proposed approach enables the agent to adapt sampling rate to video content and skip most of the frames without the loss of information. Second, in order to have a confident decision, the number of frames that should be watched by an agent varies greatly from one video to another. We incorporate an adaptive stop network to measure confidence score and generate timely trigger to stop the agent watching videos, which improves efficiency without loss of accuracy. Our approach reduces the computational cost significantly for the large-scale YouTube-8M dataset, while the accuracy remains the same.

scholarly journals Have clinicians adopted the use of brain MRI for patients with TIA and minor stroke?

Neurology ◽  
2016 ◽  
Vol 88 (3) ◽  
pp. 237-244 ◽  
Author(s):  
Seemant Chaturvedi ◽  
Susan Ofner ◽  
Fitsum Baye ◽  
Laura J. Myers ◽  
Mike Phipps ◽  
...  
Keyword(s):  
Imaging Modality ◽  
Brain Mri ◽  
National Health System ◽  
Minor Stroke ◽  
Symptom Duration ◽  
Veterans Affairs Hospitals ◽  
Chart Abstraction ◽  
History Of ◽  
Artery Disease ◽  
Facility Level

Background:Use of MRI with diffusion-weighted imaging (DWI) can identify infarcts in 30%–50% of patients with TIA. Previous guidelines have indicated that MRI-DWI is the preferred imaging modality for patients with TIA. We assessed the frequency of MRI utilization and predictors of MRI performance.Methods:A review of TIA and minor stroke patients evaluated at Veterans Affairs hospitals was conducted with regard to medical history, use of diagnostic imaging within 2 days of presentation, and in-hospital care variables. Chart abstraction was performed in a subset of hospitals to assess clinical variables not available in the administrative data.Results:A total of 7,889 patients with TIA/minor stroke were included. Overall, 6,694 patients (84.9%) had CT or MRI, with 3,396/6,694 (50.7%) having MRI. Variables that were associated with increased odds of CT performance were age >80 years, prior stroke, history of atrial fibrillation, heart failure, coronary artery disease, anxiety, and low hospital complexity, while blood pressure >140/90 mm Hg and high hospital complexity were associated with increased likelihood of MRI. Diplopia (87% had MRI, p = 0.03), neurologic consultation on the day of presentation (73% had MRI, p < 0.0001), and symptom duration of >6 hours (74% had MRI, p = 0.0009) were associated with MRI performance.Conclusions:Within a national health system, about 40% of patients with TIA/minor stroke had MRI performed within 2 days. Performance of MRI appeared to be influenced by several patient and facility-level variables, suggesting that there has been partial acceptance of the previous guideline that endorsed MRI for patients with TIA.

The Pregnant Surgical Patient

2015 ◽  
Author(s):  
Nina Tamirisa ◽  
Sami Kilic ◽  
Mostafa Borahay
Keyword(s):  
Critical Care ◽  
Abdominal Pain ◽  
Imaging Modality ◽  
Pregnant Patient ◽  
Surgical Patient ◽  
Changes In Pregnancy ◽  
Surgical Conditions ◽  
Laboratory Changes ◽  
In Pregnancy

The most vulnerable time for a fetus is during embryogenesis in the first 8 to 10 weeks of pregnancy, when women may be unaware of their pregnancy. Once pregnancy is established, a standard approach to the pregnant patient is the optimal way to ensure medical and surgical decisions are made within the context of maintaining the safety of both mother and fetus. This review describes the approach to the pregnant patient for surgical conditions within the context of physiologic changes of the patient and fetus at each trimester, anesthesia and critical care in pregnancy, imaging and drugs safe for use in pregnancy, and nongynecologic surgery in the pregnant patient and specific surgical conditions. Tables outline the classification of abortion, the assessment of pregnancy viability, physiologic changes in pregnancy, laboratory changes in pregnancy, imaging modality and radiation dose, and antibiotics and safety in pregnancy. Figures include a diagram of types of hysterectomy, respiratory changes in pregnancy, and enlargement of the uterus. Algorithms outline the approach to abdominal pain in the pregnant patient and diagnosis and management of ectopic pregnancy. This review contains 5 figures, 6 tables, and 85 references.

Retrained Classification of Tyrosinase Inhibitors and “In Silico” Potency Estimation by Using Atom-Type Linear Indices

2013 ◽  
pp. 322-427
Keyword(s):  
External Validation ◽  
Correlation Coefficients ◽  
Classification Models ◽  
Training Set ◽  
Linear Discriminant ◽  
Oecd Principles ◽  
Qsar Models ◽  
Validation Set ◽  
Global Accuracy

In this paper, the authors present an effort to increase the applicability domain (AD) by means of retraining models using a database of 701 great dissimilar molecules presenting anti-tyrosinase activity and 728 drugs with other uses. Atom-based linear indices and best subset linear discriminant analysis (LDA) were used to develop individual classification models. Eighteen individual classification-based QSAR models for the tyrosinase inhibitory activity were obtained with global accuracy varying from 88.15-91.60% in the training set and values of Matthews correlation coefficients (C) varying from 0.76-0.82. The external validation set shows globally classifications above 85.99% and 0.72 for C. All individual models were validated and fulfilled by OECD principles. A brief analysis of AD for the training set of 478 compounds and the new active compounds included in the re-training was carried out. Various assembled multiclassifier systems contained eighteen models using different selection criterions were obtained, which provide possibility of select the best strategy for particular problem. The various assembled multiclassifier systems also estimated the potency of active identified compounds. Eighteen validated potency models by OECD principles were used.

Assessment of Language Learning Strategies in Technology-Enhanced Learning Environment

2020 ◽  
pp. 83-99
Author(s):  
Wagdi Rashad Ali Bin-Hady ◽  
Abdu Al-kadi ◽  
Ali Abbas Falah Alzubi ◽  
Hassan Saleh Mahdi
Keyword(s):  
Language Learning ◽  
Learning Strategies ◽  
English Language ◽  
Cognitive Strategies ◽  
Technology Enhanced Learning ◽  
Language Learning Strategies ◽  
Efl Learners ◽  
Significant Difference ◽  
And Gender

This chapter reports on the Yemeni and Saudi EFL learners' use of language learning strategies (LLSs) in technology-mediated language learning contexts. The study examines whether nationality and gender play a significant role in using LLSs on electronic platforms. The study adopted a correlative design in which 100 Yemeni and Saudi university students were recruited to respond to an online close-ended questionnaire. Drawing on Oxford's classification of learning strategies, the findings of this study showed that metacognitive and cognitive strategies were used more frequently compared to the other LLSs. Moreover, the findings of t-test showed a significant difference in the use of LLSs attributed to nationality in favor of the Saudi learners and no significant difference in the choice of LLSs attributed to gender. The study provided some suggestions for EFL learners to benefit from technology in their English language learning.

Automatic Texture Based Classification of the Dynamics of One-Dimensional Binary Cellular Automata

2019 ◽  
Vol 8 (4) ◽  
pp. 41-61
Author(s):  
Marcelo Arbori Nogueira ◽  
Pedro Paulo Balbi de Oliveira
Keyword(s):  
Cellular Automata ◽  
Dynamic Behavior ◽  
Temporal Evolution ◽  
Computational Cost ◽  
Great Variability ◽  
Texture Descriptor ◽  
One Dimensional ◽  
Binary Images ◽  
Elementary Cellular Automata

Cellular automata present great variability in their temporal evolutions due to the number of rules and initial configurations. The possibility of automatically classifying its dynamic behavior would be of great value when studying properties of its dynamics. By counting on elementary cellular automata, and considering its temporal evolution as binary images, the authors created a texture descriptor of the images - based on the neighborhood configurations of the cells in temporal evolutions - so that it could be associated to each dynamic behavior class, following the scheme of Wolfram's classic classification. It was then possible to predict the class of rules of a temporal evolution of an elementary rule in a more effective way than others in the literature in terms of precision and computational cost. By applying the classifier to the larger neighborhood space containing 4 cells, accuracy decreased to just over 70%. However, the classifier is still able to provide some information about the dynamics of an unknown larger space with reduced computational cost.

Sign in / Sign up

Export Citation Format

Share Document