Automatic monitoring of food intake in free living conditions is still an open problem to solve. This paper presents a novel necklace-type wearable system embedded with a piezoelectric sensor to monitor ingestive behavior by detecting skin motion from the lower trachea. Detected events are incorporated for food classification. Unlike the previous state-of-the-art piezoelectric sensor based system that employs spectrogram features, we have tried to fully exploit time-domain based signals for optimal features. Through numerous evaluations on the length of a frame, we have found the best performance with a frame length of 70 samples (3.5 seconds). This demonstrates that the chewing sequence carries important information for food classification. Experimental results show the validity of the proposed algorithm for food intake detection and food classification in real-life scenarios. Our system yields an accuracy of 89.2% for food intake detection and 80.3% for food classification over 17 food categories. Additionally, our system is based on a smartphone app, which helps users live healthy by providing them with real-time feedback about their ingested food episodes and types.

One of the major causes of mortalities around the globe is lung cancer with the least chance of survival even after the diagnosis. Computer-aided detection can play an important role, especially in initial screening and thus prevent the deaths caused by lung cancer. In this paper, a novel technique for lung nodule detection, which is the primary cause of lung cancer, is proposed using convolutional neural networks. Initially, the lung volume is segmented from a CT image using optimal thresholding which is followed by image enhancement using multi-scale dot enhancement filtering. Next, lung nodule candidates are detected from an enhanced image and certain features are extracted. The extracted features belong to intensity, shape and texture class. Finally, the classification of lung nodule candidates into nodules and non-nodules is done using a convolutional neural network. The Lung Image Database Consortium (LIDC) dataset has been used to evaluate the proposed system which achieved an accuracy of 94.80% with 6.2 false positives per scan only.