Dual-Comb Gas Sensor Integrated with a Neural Network-Based Spectral Decoupling Algorithm of Overlapped Spectra for Gas Mixture Sensing.

Cross-interference among absorptions severely affects the ability to achieve accurate gas concentration retrieval through gas molecular specificity. In this study, a novel dual gas sensor was proposed to separate methane and water absorbance from the blended spectra of their mixture in the mid-infrared (MIR) band by employing a neural network algorithm. To address the scarcity of experimental data, the neural network was trained over a simulated data set constructed with the same distribution as the experimental ones. The system takes advantages of the broadband spectra to provide high-quality comb data and allows the neural network to establish an accurate spectral decoupling function. In addition, a feature absorption peak screening mechanism was proposed to achieve more accurate concentration retrieval, which avoids the prediction error introduced by interrogating the only peak of the separated spectra. The promising results of the systematic evaluation have demonstrated the feasibility of our methods in practical detections.

A Multi-Task Convolutional Neural Network for Lesion Region Segmentation and Classification of Non-Small Cell Lung Carcinoma.

Targeted therapy is an effective treatment for non-small cell lung cancer. Before treatment, pathologists need to confirm tumor morphology and type, which is time-consuming and highly repetitive. In this study, we propose a multi-task deep learning model based on a convolutional neural network for joint cancer lesion region segmentation and histological subtype classification, using magnified pathological tissue images. Firstly, we constructed a shared feature extraction channel to extract abstract information of visual space for joint segmentation and classification learning. Then, the weighted losses of segmentation and classification tasks were tuned to balance the computing bias of the multi-task model. We evaluated our model on a private in-house dataset of pathological tissue images collected from Qilu Hospital of Shandong University. The proposed approach achieved Dice similarity coefficients of 93.5% and 89.0% for segmenting squamous cell carcinoma (SCC) and adenocarcinoma (AD) specimens, respectively. In addition, the proposed method achieved an accuracy of 97.8% in classifying SCC vs. normal tissue and an accuracy of 100% in classifying AD vs. normal tissue. The experimental results demonstrated that our method outperforms other state-of-the-art methods and shows promising performance for both lesion region segmentation and subtype classification.