Multimodal Visual Concept Learning with Weakly Supervised Techniques

TL;DR

This paper introduces two weakly supervised learning techniques, FSMIL and PLMIL, that leverage textual cues to improve video concept recognition from movies and screenplays, outperforming existing methods.

Contribution

The paper presents novel weakly supervised methods extending MIL with fuzzy sets and probabilistic labels, specifically designed for complex semantic extraction from natural language.

Findings

Significant improvement over state-of-the-art weakly supervised approaches

Effective in face and action recognition tasks in movies

Utilizes semantic similarity for weak label extraction

Abstract

Despite the availability of a huge amount of video data accompanied by descriptive texts, it is not always easy to exploit the information contained in natural language in order to automatically recognize video concepts. Towards this goal, in this paper we use textual cues as means of supervision, introducing two weakly supervised techniques that extend the Multiple Instance Learning (MIL) framework: the Fuzzy Sets Multiple Instance Learning (FSMIL) and the Probabilistic Labels Multiple Instance Learning (PLMIL). The former encodes the spatio-temporal imprecision of the linguistic descriptions with Fuzzy Sets, while the latter models different interpretations of each description's semantics with Probabilistic Labels, both formulated through a convex optimization algorithm. In addition, we provide a novel technique to extract weak labels in the presence of complex semantics, that consists of semantic similarity computations. We evaluate our methods on two distinct problems, namely face and action recognition, in the challenging and realistic setting of movies accompanied by their screenplays, contained in the COGNIMUSE database. We show that, on both tasks, our method considerably outperforms a state-of-the-art weakly supervised approach, as well as other baselines.