Replication Data for: Feed-forward Neural Networks for Video Scene Classification from Statistical Features

A set of colour histograms extracted from a sitcom The Big Bang Theory primarily used for scene recognition. All files are stored in Python2 pickle format. Dictionary data are stored in ordinary python dict, numerical data use a numpy library. Each zipfile includes a single episode. frame_count.p: an information on episode length in frames splits.p: list of detected shots \d{3}.p: dict with 'scene' key. \d{3}_h.p: 180-bin Hue histogram \d{3}_s.p: 256-bin Saturation histogram \d{3}_v.p: 256-bin Value histogram \d{3}_rgb_8.p: Linearised 8*8*8 RGB histogram \d{3}_hsv_20_4_4_2x2.p: Linearised histogram with 20H*4S*4V bins of picture divided into 2x2 grid