Logo des Repositoriums

Auflistung nach Schlagwort "pose estimation"

1 - 3 von 3
  • Workshopbeitrag
    A New Software Toolset for Recording and Viewing Body Tracking Data
    (Mensch und Computer 2023 - Workshopband, 2023) Fietkau, Julian
    While 3D body tracking data has been used in empirical HCI studies for many years now, the tools to interact with it tend to be either vendor-specific proprietary monoliths or single-use tools built for one specific experiment and then discarded. In this paper, we present our new toolset for cross-vendor body tracking data recording, storing, and visualization/playback. Our goal is to evolve it into an open data format along with software tools capable of producing and consuming body tracking recordings in said new format, and we hope to find interested collaborators for this endeavour.
  • Textdokument
    Deep Convolutional Neural Networks for Pose Estimation in Image-Graphics Search
    (INFORMATIK 2017, 2017) Eberts, Markus; Ulges, Adrian
    Deep Convolutional Neural Networks (CNNs) have recently been highly successful in various image understanding tasks, ranging from object category recognition over image classification to scene segmentation. We employ CNNs for pose estimation in a cross-modal retrieval system, which -given a photo of an object -allows users to retrieve the best match from a repository of 3D models. As our system is supposed to display retrieved 3D models from the same perspective as the query image (potentially with virtual objects blended over), the pose of the object relative to the camera needs to be estimated. To do so, we study two CNN models. The first is based on end-to-end learning, i.e. a regression neural network directly estimates the pose. The second uses transfer learning with a very deep CNN pre-trained on a large-scale image collection. In quantitative experiments on a set of 3D models and real-world photos of chairs, we compare both models and show that while the end-to-end learning approach performs well on the domain it was trained on (graphics) it suffers from the capability to generalize to a new domain (photos). The transfer learning approach on the other hand handles this domain drift much better, resulting in an average angle deviation from the ground truth angle of about 14 degrees on photos.
  • Konferenzbeitrag
    Region-Based CNNs for Pedestrian Gender Recognition in Visual Surveillance Environments
    (BIOSIG 2019 - Proceedings of the 18th International Conference of the Biometrics Special Interest Group, 2019) Yaghoubi, Ehsan; Alirezazadeh, Pendar; Assunção, Eduardo; Neves, João C.; Proença, Hugo
    Inferring soft biometric labels in totally uncontrolled outdoor environments, such as surveillance scenarios, remains a challenge due to the low resolution of data and its covariates that might seriously compromise performance (e.g., occlusions and subjects pose). In this kind of data, even state-of-the-art deep-learning frameworks (such as ResNet) working in a holistic way, attain relatively poor performance, which was the main motivation for the work described in this paper. In particular, having noticed the main effect of the subjects’ “pose” factor, in this paper we describe a method that uses the body keypoints to estimate the subjects pose and define a set of regions of interest (e.g., head, torso, and legs). This information is used to learn appropriate classification models, specialized in different poses/body parts, which contributes to solid improvements in performance. This conclusion is supported by the experiments we conducted in multiple real-world outdoor scenarios, using the data acquired from advertising panels placed in crowded urban environments.