Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

General General

Learning image features with fewer labels using a semi-supervised deep convolutional network.

In Neural networks : the official journal of the International Neural Network Society

Learning feature embeddings for pattern recognition is a relevant task for many applications. Deep learning methods such as convolutional neural networks can be employed for this assignment with different training strategies: leveraging pre-trained models as baselines; training from scratch with the target dataset; or fine-tuning from the pre-trained model. Although there are separate systems used for learning features from labelled and unlabelled data, there are few models combining all available information. Therefore, in this paper, we present a novel semi-supervised deep network training strategy that comprises a convolutional network and an autoencoder using a joint classification and reconstruction loss function. We show our network improves the learned feature embedding when including the unlabelled data in the training process. The results using the feature embedding obtained by our network achieve better classification accuracy when compared with competing methods, as well as offering good generalisation in the context of transfer learning. Furthermore, the proposed network ensemble and loss function is highly extensible and applicable in many recognition tasks.

Dos Santos Fernando P, Zor Cemre, Kittler Josef, Ponti Moacir A

2020-Aug-25

Feature generalisation, Semi-supervised learning, Transfer learning

Ophthalmology Ophthalmology

Toward automated severe pharyngitis detection with smartphone camera using deep learning networks.

In Computers in biology and medicine

PURPOSE : Severe pharyngitis is frequently associated with inflammations caused by streptococcal pharyngitis, which can cause immune-mediated and post-infectious complications. The recent global pandemic of coronavirus disease (COVID-19) encourages the use of telemedicine for patients with respiratory symptoms. This study, therefore, purposes automated detection of severe pharyngitis using a deep learning framework with self-taken throat images.

METHODS : A dataset composed of two classes of 131 throat images with pharyngitis and 208 normal throat images was collected. Before the training classifier, we constructed a cycle consistency generative adversarial network (CycleGAN) to augment the training dataset. The ResNet50, Inception-v3, and MobileNet-v2 architectures were trained with transfer learning and validated using a randomly selected test dataset. The performance of the models was evaluated based on the accuracy and area under the receiver operating characteristic curve (ROC-AUC).

RESULTS : The CycleGAN-based synthetic images reflected the pragmatic characteristic features of pharyngitis. Using the synthetic throat images, the deep learning model demonstrated a significant improvement in the accuracy of the pharyngitis diagnosis. ResNet50 with GAN-based augmentation showed the best ROC-AUC of 0.988 for pharyngitis detection in the test dataset. In the 4-fold cross-validation using the ResNet50, the highest detection accuracy and ROC-AUC achieved were 95.3% and 0.992, respectively.

CONCLUSION : The deep learning model for smartphone-based pharyngitis screening allows fast identification of severe pharyngitis with a potential of the timely diagnosis of pharyngitis. In the recent pandemic of COVID-19, this framework will help patients with upper respiratory symptoms to improve convenience in diagnosis and reduce transmission.

Yoo Tae Keun, Choi Joon Yul, Jang Younil, Oh Ein, Ryu Ik Hee

2020-Aug-20

Automated diagnosis, Deep learning, Pharyngitis, Smartphone, Telemedicine, Tonsillitis

Radiology Radiology

Radiomic signature-based nomogram to predict disease-free survival in stage II and III colon cancer.

In European journal of radiology ; h5-index 47.0

PURPOSE : To develop a radiomic nomogram to predict disease-free survival (DFS) in patients with colon cancer.

METHODS : We retrospectively identified 302 patients with stage III colon cancer and 269 patients with stage II colon cancer who had undergone multidetector computed tomography (MDCT) and radical resection between January 2009 and December 2015. Patients were divided into a training cohort (n = 322) and an external validation cohort (n = 249). Radiomic features were extracted from MDCT images, and a radiomic signature was built as to predict DFS. A radiomic nomogram integrating the radiomic signature and clinicopathologic characteristics was developed using multivariable logistic regression. The nomogram was evaluated with regard to calibration, discrimination, and clinical utility.

RESULTS : The radiomic signature was an independent prognostic factor for DFS in the training cohort (HR = 1.102; 95 % CI: 1.052-1.156; P < 0.001) and the external validation cohort (HR = 1.157; 95 % CI: 1.030-1.301; P = 0.014). The radiomic signature-based nomogram was more effective at predicting DFS than either the TNM staging system or a clinicopathologic nomogram. The C-indices of the radiomic nomogram and TNM staging system were 0.780 (95 % CI: 0.734-0.847) and 0.738 (0.687-0.784) respectively. The radiomic signature-based nomogram demonstrated good fitness (shown by calibration curves) and clinical usefulness (shown by decision curve analysis).

CONCLUSION : A radiomic signature derived from MDCT images can effectively predict DFS in patients with stage II and III colon cancer and could be used as a supplement for risk stratification.

Yao Xun, Sun Caixia, Xiong Fei, Zhang Xinyu, Cheng Jin, Wang Chao, Ye Yingjiang, Hong Nan, Wang Lihui, Liu Zhenyu, Meng Xiaochun, Wang Yi, Tian Jie

2020-Aug-19

Colon cancer, Computed tomography, Disease-free survival, Radiomics

General General

Pain phenotypes classified by machine learning using electroencephalography features.

In NeuroImage ; h5-index 117.0

Pain is a multidimensional experience mediated by distributed neural networks in the brain. To study this phenomenon, EEGs were collected from 20 subjects with chronic lumbar radiculopathy, 20 age and gender matched healthy subjects, and 17 subjects with chronic lumbar pain scheduled to receive an implanted spinal cord stimulator. Analysis of power spectral density, coherence, and phase-amplitude coupling using conventional statistics showed that there were no significant differences between the radiculopathy and control groups after correcting for multiple comparisons. However, analysis of transient spectral events showed that there were differences between these two groups in terms of the number, power, and frequency-span of events in a low gamma band. Finally, we trained a binary support vector machine to classify radiculopathy versus healthy subjects, as well as a 3-way classifier for subjects in the 3 groups. Both classifiers performed significantly better than chance, indicating that EEG features contain relevant information pertaining to sensory states, and may be used to help distinguish between pain states when other clinical signs are inconclusive.

Levitt Joshua, Edhi Muhammad M, Thorpe Ryan V, Leung Jason W, Michishita Mai, Koyama Suguru, Yoshikawa Satoru, Scarfo Keith A, Carayannopoulos Alexios G, Gu Wendy, Srivastava Kyle H, Clark Bryan A, Esteller Rosana, Borton David A, Jones Stephanie R, Saab Carl Y

2020-Aug-29

General General

Gait Phase Recognition Using Deep Convolutional Neural Network with Inertial Measurement Units.

In Biosensors

Gait phase recognition is of great importance in the development of assistance-as-needed robotic devices, such as exoskeletons. In order for a powered exoskeleton with phase-based control to determine and provide proper assistance to the wearer during gait, the user's current gait phase must first be identified accurately. Gait phase recognition can potentially be achieved through input from wearable sensors. Deep convolutional neural networks (DCNN) is a machine learning approach that is widely used in image recognition. User kinematics, measured from inertial measurement unit (IMU) output, can be considered as an 'image' since it exhibits some local 'spatial' pattern when the sensor data is arranged in sequence. We propose a specialized DCNN to distinguish five phases in a gait cycle, based on IMU data and classified with foot switch information. The DCNN showed approximately 97% accuracy during an offline evaluation of gait phase recognition. Accuracy was highest in the swing phase and lowest in terminal stance.

Su Binbin, Smith Christian, Gutierrez Farewik Elena

2020-Aug-27

IMU, convolutional neural network, gait phase recognition

General General

Identifying policy challenges of COVID-19 in hardly reliable data and judging the success of lockdown measures.

In Journal of population economics

Identifying structural breaks in the dynamics of COVID-19 contagion is crucial to promptly assess policies and evaluate the effectiveness of lockdown measures. However, official data record infections after a critical and unpredictable delay. Moreover, people react to the health risks of the virus and also anticipate lockdowns. All of this makes it complex to quickly and accurately detect changing patterns in the virus's infection dynamic. We propose a machine learning procedure to identify structural breaks in the time series of COVID-19 cases. We consider the case of Italy, an early-affected country that was unprepared for the situation, and detect the dates of structural breaks induced by three national lockdowns so as to evaluate their effects and identify some related policy issues. The strong but significantly delayed effect of the first lockdown suggests a relevant announcement effect. In contrast, the last lockdown had significantly less impact. The proposed methodology is robust as a real-time procedure for early detection of the structural breaks: the impact of the first two lockdowns could have been correctly identified just the day after they actually occurred.

Bonacini Luca, Gallo Giovanni, Patriarca Fabrizio

2020-Aug-26

COVID-19, Coronavirus, Lockdown, Machine learning