VSD, a public dataset for the detection of violent scenes in movies: design, annotation, analysis and evaluation

Abstract

International audienceContent-based analysis to find where violence appears in multimedia content has several applications, from parental control and children protection to surveillance. This paper presents the design and annotation of the Violent Scene Detection dataset, a corpus targeting the detection of physical violence in Hollywood movies. We discuss definitions of physical violence and provide a simple and objective definition which was used to annotate a set of 18 movies, thus resulting in the largest freely-available dataset for such a task. We discuss borderline cases and compare with annotations based on a subjective definition which requires multiple annotators. We provide a detailed analysis of the corpus, in particular regarding the relationship between violence and a set of key audio and visual concepts which were also annotated. The VSD dataset results from two years of benchmarking in the framework of the MediaEval initiative. We provide results from the 2011 and 2012 benchmarks as a validation of the dataset and as a state-of-the-art baseline. The VSD dataset is freely available at the address: http://www.technicolor.com/en/innovation/research-innovation/scientific-data-sharing/violent-scenes-dataset.

Sử Dụng Deep Neural Networks Biểu Diễn Các Thuộc Tính Cho Bài Toán Phát Hiện Cảnh Bạo Lực Trong Video

“…Đầu vào của hệ thống là các video mà cụ thể ở đây trong dữ liệu mà chúng tôi sử dụng từ cuộc thi MediaEval Affect Task [17][18] là các bộ phim Hollywood. Các video sẽ được cắt thành các đoạn (shot) mỗi đoạn có thời lượng là 5 giây, trong mỗi đoạn chúng tôi sẽ lấy mẫu theo tần suất 5 cảnh (keyframe)/ giây được làm dữ liệu đầu vào cho quá trình rút trích đặc trưng tiếp theo.…”

A modified vector of locally aggregated descriptors approach for fast video classification

“…This dataset (in its various versions) has been exploited during the 2011-2014 MediaEval benchmarking campaigns [10];…”

“…Sample images from the experimentation datasets: Blip10000 [52], ADL [36], VSD2013 [10] and UCF101 [26] …”

“…We also test our approach on the 2013 Violence Scenes Detection VSD2013 dataset [10]. This data is annotated at video shot level for the presence of violence (yes/no annotations).…”

“…Comparison with state-of-the-art (VSD2013[10] shot based classification; the best results are presented in bold)…”