One-Shot-Learning Gesture Recognition using HOG-HOF Features

TL;DR

This paper introduces a one-shot gesture recognition system using combined HOG and HOF features on RGB and depth data, employing novel histogram comparison and video trimming techniques to improve accuracy.

Contribution

It presents a new approach combining appearance and motion descriptors with a novel video trimming algorithm for improved gesture recognition.

Findings

Methods outperform existing published approaches

Significant narrowing of the gap between human and algorithm performance

Effective use of cross-bin histogram relationships

Abstract

The purpose of this paper is to describe one-shot-learning gesture recognition systems developed on the \textit{ChaLearn Gesture Dataset}. We use RGB and depth images and combine appearance (Histograms of Oriented Gradients) and motion descriptors (Histogram of Optical Flow) for parallel temporal segmentation and recognition. The Quadratic-Chi distance family is used to measure differences between histograms to capture cross-bin relationships. We also propose a new algorithm for trimming videos --- to remove all the unimportant frames from videos. We present two methods that use combination of HOG-HOF descriptors together with variants of Dynamic Time Warping technique. Both methods outperform other published methods and help narrow down the gap between human performance and algorithms on this task. The code has been made publicly available in the MLOSS repository.