Data-efficient resolution transfer with Continuous Kernel Convolutional Neural Networks

Convolutional Neural Networks (CNNs) benefit from fine-grained details in high-resolution images, but these images are not always easily available as data collection can be expensive or time-consuming. Transfer learning pre-trains models on data from a related domain before fine-tuning on the main domain, and is a common strategy to deal with...

Frequency learning for structured CNN filters with Gaussian fractional derivatives

A structured CNN filter basis allows incorporating priors about natural image statistics and thus require less training examples to learn, saving valuable annotation time. Here, we build on the Gaussian derivative CNN filter basis that learn both the orientation and scale of the filters. However, this Gaussian filter basis definition depends on...

Contrastive Learning of Visual Representations from Unlabeled Videos

This thesis presents a novel self-supervised approach of learning visual representations from videos containing human actions. Our approach tackles the complex problem of learning without the need of labeled data by exploring to what extent the ideas successfully used for images can be transferred, adapted and extended to videos for action...

Trunk-Branch Ensemble Convolutional Neural Networks for Large Scale, Few-Shot Video-to-Still Face Recognition

For the real-world face recognition, factors such as occlusion and pose-variant (cross face) would affect the identification/verification performance. In addition, large number of classes also increase the complexity, which makes verification/identification even harder. In order to deal with these issues, how to extract discriminative embeddings...