Sorry, you need to enable JavaScript to visit this website.

We examine two key questions in GAN training, namely overfitting and mode drop, from an empirical perspective. We show that when stochasticity is removed from the training procedure, GANs can overfit and exhibit almost no mode drop. Our results shed light on important characteristics of the GAN training procedure. They also provide evidence against prevailing intuitions that GANs do not memorize the training set, and that mode dropping is mainly due to properties of the GAN objective rather than how it is optimized during training.

Deception detection has been a hot research topic in many areas such as jurisprudence, law enforcement, business, and computer vision. However, there are still many problems that are worth more investigation. One of the major challenges is the data scarcity problem. So far, only one multi-modal benchmark dataset on deception detection has been published, which contains 121 video clips for deception detection (61 for deceptive class and 60 for truthful class).

We propose a frontalization technique for 2D facial land- marks, designed to aid in the analysis of facial expressions. It employs a new normalization strategy aiming to minimize identity variations, by displacing groups of facial landmarks to standardized locations. The technique operates directly on 2D landmark coordinates, does not require additional feature extraction and as such is computationally light. It achieves considerable improvement over a reference approach, justifying its use as an efficient preprocessing step for facial expression analysis based on geometric features.

Mobile devices are the source of a vast majority of digital photos today. Photos taken by mobile devices generally have fairly good visual quality. When evaluating high-quality mobile device photos, people have to manually zoom in to local regions to discern the subtle difference. Understandably, a global objective quality assessment method cannot perform well on such task. Therefore, local region selection is widely recognized as a prerequisite for the following quality evaluation.

Many objective video quality assessment (VQA) algorithms include a key step of temporal pooling of frame-level quality scores. However, less attention has been paid to studying the relative efficiencies of different pooling methods on no-reference (blind) VQA. Here we conduct a large-scale comparative evaluation to assess the capabilities and limitations of multiple temporal pooling strategies on blind VQA of user-generated videos. The study yields insights and general guidance regarding the application and selection of temporal pooling models.

Saliency has been widely studied in relation to image quality assessment (IQA). The optimal use of saliency in IQA metrics, however, is nontrivial and largely depends on whether saliency can be accurately predicted for images containing various distortions. Although tremendous progress has been made in saliency modelling, very little is known about whether and to what extent state-of-the-art methods are beneficial for saliency prediction of distorted images.

Advances in federated learning and edge computing advocate for deep learning models to run at edge devices for video analysis. However, the captured video frame rate is too high to be processed at the edge in real-time with a typical model such as CNN. Any approach to consecutively feed frames to the model compromises both the quality (by missing important frames) and the efficiency (by processing redundantly similar frames) of analysis.