Several leading depth-from-stereo methods rely on intensive 4D cost volumes and computationally intensive 3D convolutions for feature matching. We propose to integrate two independent concepts into a new 4D cost volume to achieve a symbiotic relationship. A feature matching part is responsible for identifying matching pixel pairs along the baseline, while a concurrent image volume part is inspired by depth-from-mono CNNs. More technically, the processing of the 4D cost volume is divided into a 2D propagation and a 3D propagation part. Starting from the feature maps of the left image, the 2D propagation, instead of directly predicting depth, supports the 3D propagation part of the cost volume at different levels by adding visual features to the geometric context. Combining the two parts reduces the number of the 3D convolution layers without sacrificing accuracy. Experiments show that our end-to-end trained CNN ranks 2nd on the KITTI2012 and ETH3D benchmarks, while being significantly faster than the 1st ranked method. The source code is available at https://github.com/ohkwon718/icvp.

Forecasting human motion of multiple persons is very challenging. It requires to model the interactions between humans and the interactions with objects and the environment. For example, a person might want to make a coffee, but if the coffee machine is already occupied the person will haveto wait. These complex relations between scene geometry and persons ariseconstantly in our daily lives, and models that wish to accurately forecasthuman behavior will have to take them into consideration. To facilitate research in this direction, we propose Humans in Kitchens, alarge-scale multi-person human motion dataset with annotated 3D human poses, scene geometry and activities per person and frame.Our dataset consists of over 7.3h recorded data of up to 16 persons at the same time in four kitchen scenes, with more than 4M annotated human poses, represented by a parametric 3D body model. In addition, dynamic scene geometry and objects like chair or cupboard are annotated per frame. As first benchmarks, we propose two protocols for short-term and long-term human motion forecasting.

A horseback riding simulator is developed to overcome challenges occurring on real horses. The goal of the simulator is to provide a real horseback riding-like sensation to the user in various situations with limited workspace. This paper focuses on generating various trajectories at different walking speeds with dynamic effects. A walking horse dynamic model is purposed based on a two-linked rimless wheel and is modified with 3D-captured data from a real horse. A controller is also designed for passing and compression-recoil modes, aiding in changing the models walking speed using speed change input. A dynamics simulation is performed and results show that the suggested walking model covers trajectory with measured data and shows more variation.

Because declining rates of population growth have left aging societies in many countries around the world, Smart Walker systems were created to assist elderly people in managing their physical, cognitive, and emotional challenges. While the main application of the Smart Walker is to assist in the daily life of the elderly, it may also be used to assist in rehabilitation. In this paper, a no-contact system is proposed to motivate exercise on a Smart Walker. A tablet device is attached to the chair in front of the user to serve as a channel for interaction. The tablet device has several modalities for expression including a 3D robot-face character, TTS (Text to Speech), and dialog messaging. The exercise motivational system selects appropriate expressions to display on the tablet device, from pre-recorded behavioral data, using real-time data from the movements of the patient.

As the number of mobile robot applications increases, research related to physical contact between humans and robots is becoming an important issues in the service robot domain. In this paper, a mobile robot platform based on spring loaded casters is proposed as solution to problems in the service robot domain. A simple dynamic equation in two dimensions was computed and the dynamic characteristics according to variation of the elastic coefficient were analyzed. In order to detect force from a human, a linear equation was formulated by regression of the driving data. With the estimated force, an impedance model is applied to the controller for compliance. Through this process, it is shown that the proposed mobile robot platform is capable of a compliant reaction to external force through of its posture, similar to unstable mobile platform, though the mobile platform has the static stability.