Velichko, Andrei; Huyut, Mehmet Tahir

Respiratory activity is a direct and interpretable physiological channel for wearable stress and affective-state recognition, yet many studies emphasize classification accuracy without identifying which respiratory properties separate different states. This work reframes RESP-based recognition as a joint predictive and explanatory problem. Using the chest respiratory channel of the WESAD dataset, we analyze 60 s windows under leave-one-subject-out validation and combine two complementary branches: compact raw-signal one-dimensional convolutional neural networks (1D-CNNs) and physically grouped handcrafted respiratory signatures. The primary application task is binary stress versus non-stress detection, while baseline, stress, amusement, and meditation are additionally analyzed in a one-vs-rest setting to reveal state-specific respiratory markers. The feature space is organized into respiratory timing, breath-to-breath variability, waveform statistics, spectral/time–frequency descriptors, and autocorrelation/nonlinear predictability descriptors, with the raw 60 s signal treated as a sixth representation for the CNN branch. We introduce autocorrelation transition lags (Zpm/Zmp) as interpretable markers of respiratory correlation scale and separately evaluate exploratory forecast-error-growth/Lyapunov-like descriptors. In the final CNN refit setting, the raw-signal model achieved the strongest stress-vs-rest performance (accuracy 96.72%, macro-F1 95.30%, MCC 90.61%). In contrast, compact feature models were stronger for baseline (MCC 65.34%), amusement (MCC 35.69%), and especially meditation (MCC 88.65%). Strict nested reanalysis of the auxiliary Top-20 stress search yielded mean MCC=82.29% versus 85.45% originally, indicating modest selection optimism. These results show that CNNs are most useful for the practical stress detector, whereas interpretable respiratory signatures provide stronger and more physiologically transparent state-specific markers for several non-stress conditions.

Alici-Karaca, Demet; Akay, Bahriye; Yay, Arzu; Suna, Pinar; Nalbantoglu, O. Ufuk; Karaboga, Dervis; Basturk, Alper; Balcioglu, Esra; Baran, Munevver

Histopathological image analysis is used in the diagnosis of many diseases such as cancer, brain tumor, fatty liver and congenital heart diseases. In clinical practice, the time-consuming diagnostic process requires an expertise in the field, and disagreements can arise among pathologists at the decision stage. Machine learning methods can be used in diagnostic stage to overcome these issues. Especially, the use of deep learning systems in the analysis of histopathological images can reduce the workload of pathologists and can provide more objective results. Over traditional machine learning methods using handcrafted features, deep learning approaches that can automatically learn features stand out with their performance in the analysis of histopathology images. In this paper, we propose a patch-based lightweight convolution neural network for histopathological image classification for detecting radiation-induced liver damage. The proposed model was trained to classify the radiation-induced liver disease dataset comprising 555 histopathological images. An accuracy of 100% for binary classification and an accuracy of 87.57% for multi-class classification were achieved on the unseen test set. Our model outperforms state-of-the-art models, including ResNet-50, AlexNet, Vgg16, and GoogleNet. This paper is the first to classify radiation-induced liver disease data with deep learning, and it can be a promising tool for diagnosis purpose of pathologists. Since it is vital to detect a damage that leads to deteriorating liver function, this study is valuable in the field of medicine.