Pavlo Molchanov

Hand gesture recognition with 3D convolutional neural networks

Touchless hand gesture recognition systems are becoming important in automotive user interfaces as they improve safety and comfort. Various computer vision algorithms have employed color and depth cameras for hand gesture recognition, but robust classification of gestures from different subjects performed under widely varying lighting conditions is still challenging. We propose an algorithm for drivers hand gesture recognition from challenging depth and intensity data using 3D convolutional neural networks. Our solution combines information from multiple spatial scales for the final prediction. It also employs spatio-temporal data augmentation for more effective training and to reduce potential overfitting. Our method achieves a correct classification rate of 77.5% on the VIVA challenge dataset.

Online Detection and Classification of Dynamic Hand Gestures with Recurrent 3D Convolutional Neural Networks

Automatic detection and classification of dynamic hand gestures in real-world systems intended for human computer interaction is challenging as: 1) there is a large diversity in how people perform gestures, making detection and classification difficult, 2) the system must work online in order to avoid noticeable lag between performing a gesture and its classification, in fact, a negative lag (classification before the gesture is finished) is desirable, as feedback to the user can then be truly instantaneous. In this paper, we address these challenges with a recurrent three-dimensional convolutional neural network that performs simultaneous detection and classification of dynamic hand gestures from multi-modal data. We employ connectionist temporal classification to train the network to predict class labels from inprogress gestures in unsegmented input streams. In order to validate our method, we introduce a new challenging multimodal dynamic hand gesture dataset captured with depth, color and stereo-IR sensors. On this challenging dataset, our gesture recognition system achieves an accuracy of 83:8%, outperforms competing state-of-the-art algorithms, and approaches human accuracy of 88:4%. Moreover, our method achieves state-of-the-art performance on SKIG and ChaLearn2014 benchmarks.

Multi-sensor system for driver's hand-gesture recognition

We propose a novel multi-sensor system for accurate and power-efficient dynamic car-driver hand-gesture recognition, using a short-range radar, a color camera, and a depth camera, which together make the system robust against variable lighting conditions. We present a procedure to jointly calibrate the radar and depth sensors. We employ convolutional deep neural networks to fuse data from multiple sensors and to classify the gestures. Our algorithm accurately recognizes 10 different gestures acquired indoors and outdoors in a car during the day and at night. It consumes significantly less power than purely vision-based systems.

Short-range FMCW monopulse radar for hand-gesture sensing

Intelligent driver assistance systems have become important in the automotive industry. One key element of such systems is a smart user interface that tracks and recognizes drivers hand gestures. Hand gesture sensing using traditional computer vision techniques is challenging because of wide variations in lighting conditions, e.g. inside a car. A short-range radar device can provide additional information, including the location and instantaneous radial velocity of moving objects. We describe a novel end-to-end (hardware, interface, and software) short-range FMCW radar-based system designed to effectively sense dynamic hand gestures. We provide an effective method for selecting the parameters of the FMCW waveform and for jointly calibrating the radar system with a depth sensor. Finally, we demonstrate that our system guarantees reliable and robust performance.

Multilayer and Multimodal Fusion of Deep Neural Networks for Video Classification

This paper presents a novel framework to combine multiple layers and modalities of deep neural networks for video classification. We first propose a multilayer strategy to simultaneously capture a variety of levels of abstraction and invariance in a network, where the convolutional and fully connected layers are effectively represented by our proposed feature aggregation methods. We further introduce a multimodal scheme that includes four highly complementary modalities to extract diverse static and dynamic cues at multiple temporal scales. In particular, for modeling the long-term temporal information, we propose a new structure, FC-RNN, to effectively transform pre-trained fully connected layers into recurrent layers. A robust boosting model is then introduced to optimize the fusion of multiple layers and modalities in a unified way. In the extensive experiments, we achieve state-of-the-art results on two public benchmark datasets: UCF101 and HMDB51.

Towards selecting robust hand gestures for automotive interfaces

Driver distraction is a serious threat to automotive safety. The visual-manual interfaces in cars are a source of distraction for drivers. Automotive touch-less hand gesture-based user interfaces can help to reduce driver distraction and enhance safety and comfort. The choice of hand gestures in automotive interfaces is central to their success and widespread adoption. In this work we evaluate the recognition accuracy of 25 different gestures for state-of-the-art computer vision-based gesture recognition algorithms and for human observers. We show that some gestures are consistently recognized more accurately than others by both vision-based algorithms and humans. We further identify similarities in the hand gesture recognition abilities of vision-based systems and humans. Lastly, by merging pairs of gestures with high miss-classification rates, we propose ten robust hand gestures for automotive interfaces, which are classified with high and equal accuracy by vision-based algorithms.

Hand Pose Estimation via Latent 2.5D Heatmap Regression

Estimating the 3D pose of a hand is an essential part of human-computer interaction. Estimating 3D pose using depth or multi-view sensors has become easier with recent advances in computer vision, however, regressing pose from a single RGB image is much less straightforward. The main difficulty arises from the fact that 3D pose requires some form of depth estimates, which are ambiguous given only an RGB image. In this paper we propose a new method for 3D hand pose estimation from a monocular image through a novel 2.5D pose representation. Our new representation estimates pose up to a scaling factor, which can be estimated additionally if a prior of the hand size is given. We implicitly learn depth maps and heatmap distributions with a novel CNN architecture. Our system achieves state-of-the-art accuracy for 2D and 3D hand pose estimation on several challenging datasets in presence of severe occlusions.