The advancement in deep learning has played an important role in revolutionizing computing and pattern recognition especially in the field of image processing. Deep learning algorithms are widely applied in the fields of image and video processing, medical imaging, self-driving cars, networks, stock market prediction, and many more application domains. These deep learning models perform better compared to traditional models due to their ability to use large amounts of data. As the data increases, the performance of deep learning models continues to improve compared to traditional machine learning models. This is the reason deep learning models have proven to be successful for big data.

However, the excellent performance of these deep learning models is generally based on labeled data. Therefore, the performance is highly dependent on labeled datasets. Dataset labeling is a time-consuming and tedious task for domain experts. The issue is even more challenging for the medical imaging domain as the dataset needs very precise labeling from physicians. The unavailability of such a large amount of labeled data causes serious problems for deep learning models such as overfitting. Researchers have used semi-supervised approaches to handle this issue. Recently, techniques have been developed to handle such scenarios where the unavailability of a large amount of data is handled efficiently. The techniques like You Only Look Once (YOLO), few short learning, zero short learning, and GANs are trying to handle this issue.

This Special Issue will focus on novel, intelligent, and efficient algorithms to handle the unavailability of labeled data using the latest algorithms for medical imaging, semantic segmentation, image classification, video summarization, video surveillance, action recognition, gesture recognition, and activity monitoring. The aim is to propose methods that can handle big data with limited labels. Submissions should be more application-oriented in the field of image processing, video processing, and medical imaging. Original research and review articles are welcome.

Potential topics include but are not limited to the following:

• Semi-supervised data labeling
• Medical imaging classification and segmentation
• Few/zero short learning
• Event recognition techniques in video and speech
• Semi-supervised techniques for pseudo label generation
• Data augmentation for active learning
• Transfer learning in semantic segmentation
• Video surveillance and activity recognition
• Online predictive modeling