Various imaging techniques combined with machine learning (ML) models have been used to build computer-aided diagnosis (CAD) systems for breast cancer (BC) detection and classification. The rise of deep learning models in recent years, represented by convolutional neural network (CNN) models, has pushed the accuracy of ML-based CAD systems to a new level that is comparable to human experts. Existing studies have explored the usage of a wide spectrum of CNN models for BC detection, and supervised learning has been the mainstream. In this study, we propose a semi-supervised learning framework based on the Vision Transformer (ViT). The ViT is a model that has been validated to outperform CNN models on numerous classification benchmarks but its application in BC detection has been rare. The proposed method offers a custom semi-supervised learning procedure that unifies both supervised and consistency training to enhance the robustness of the model. In addition, the method uses an adaptive token sampling technique that can strategically sample the most significant tokens from the input image, leading to an effective performance gain. We validate our method on two datasets with ultrasound and histopathology images. Results demonstrate that our method can consistently outperform the CNN baselines for both learning tasks. The code repository of the project is available at https://github.com/FeiYee/Breast-area-TWO.

The performance of fruit surface defect detection is easily affected by factors such as noisy background and foliage occlusion. In this study, we choose lychee as a fruit type to investigate its surface quality. Lychees are hard to preserve and have to be stored at low temperatures to keep fresh. Additionally, the surface of lychees is subject to scratches and cracks during harvesting/processing. To explore the feasibility of the automation of defective surface detection for lychees, we build a dataset with 3743 samples divided into three categories, namely, mature, defects, and rot. The original dataset suffers an imbalanced distribution issue. To address it, we adopt a transformer-based generative adversarial network (GAN) as a means of data augmentation that can effectively enhance the original training set with more and diverse samples to rebalance the three categories. In addition, we investigate three deep convolutional neural network (DCNN) models, including SSD-MobileNet V2, Faster RCNN-ResNet50, and Faster RCNN-Inception-ResNet V2, trained under different settings for an extensive comparison study. The results show that all three models demonstrate consistent performance gains in mean average precision (mAP), with the application of GAN-based augmentation. The rebalanced dataset also reduces the inter-category discrepancy, allowing a DCNN model to be trained equally across categories. In addition, the qualitative results show that models trained under the augmented setting can better identify the critical regions and the object boundary, leading to gains in mAP. Lastly, we conclude that the most cost-effective model, SSD-MobileNet V2, presents a comparable mAP (91.81%) and a superior inference speed (102 FPS), suitable for real-time detection in industrial-level applications.

Food defect detection is crucial for the automation of food production and processing. Potato surface defect detection remains challenging due to the irregular shape of potato individuals and various types of defects. This paper employs deep convolutional neural network (DCNN) models for potato surface defect detection. In particular, we applied transfer learning by fine-tuning a base model through three DCNN models—SSD Inception V2, RFCN ResNet101, and Faster RCNN ResNet101—on a self-developed dataset, and achieved an accuracy of 92.5%, 95.6%, and 98.7%, respectively. RFCN ResNet101 presented the best overall performance in detection speed and accuracy. It was selected as the final model for out-of-sample testing, further demonstrating the model’s ability to generalize.

Building an accurate and robust energy consumption prediction model is a core mission for the energy management and operation system of a smart building. Prior efforts have explored numerous predictive models in various load prediction scenarios. However, the joint effect of data augmentation and ensemble learning in energy forecasting has not been fully explored. In this paper, we propose a generative adversarial network (GAN)-enhanced ensemble model for energy consumption forecasting in large commercial buildings. The ensemble model aggregates five constituent models with a stacking ensemble method. In addition, we employ a GAN to learn the sample distribution from the original dataset and generate high-quality samples to enhance the training set. The augmented dataset allows a model to be trained with more diverse samples to increase its robustness. A series of experiments are conducted to validate the proposed method with three GAN variants using three performance metrics, including mean absolute error (MAE), root mean square error (RMSE), and coefficient of variation of root mean square error (CVRMSE). Results demonstrate the GAN-enhanced ensemble models are more robust with consistent improvement in reducing the prediction errors. The best ensemble model, enhanced by the Information Maximizing GAN (InfoGAN), outperforms the model without augmentation by decreasing the average error by 1.71, 1.63, and 4.72%, in MAE, RMSE, and CVRMSE respectively, validating its piratical value for building an energy consumption forecasting model in a real-world system.

Medical images of brain tumors are critical for characterizing the pathology of tumors and early diagnosis. There are multiple modalities for medical images of brain tumors. Fusing the unique features of each modality of the magnetic resonance imaging (MRI) scans can accurately determine the nature of brain tumors. The current genetic analysis approach is time-consuming and requires surgical extraction of brain tissue samples. Accurate classification of multi-modal brain tumor images can speed up the detection process and alleviate patient suffering. Medical image fusion refers to effectively merging the significant information of multiple source images of the same tissue into one image, which will carry abundant information for diagnosis. This paper proposes a novel attentive deep-learning-based classification model that integrates multi-modal feature aggregation, lite attention mechanism, separable embedding, and modal-wise shortcuts for performance improvement. We evaluate our model on the RSNA-MICCAI dataset, a scenario-specific medical image dataset, and demonstrate that the proposed method outperforms the state-of-the-art (SOTA) by around 3%.