Biology subjects

Labuguen, R.

Publications and source records attributed to Labuguen, R..

4 recordsLinked to original sources

bioRxiv2021

Multiple Monkey Pose Estimation Using OpenPose

This paper proposes a system for pose estimation on monkeys, in diverse and challenging scenarios, and under complex social interactions by using OpenPose. In comparison to most animals used for research, Monkeys present additional difficulties for pose estimation. Multiple degrees of freedom, unique complex postures, intricated social interactions, among others. Our monkey OpenPose trained model is robust against these difficulties. It achieves similar performance as in human pose estimation models, and it can run in Realtime.

neuroscience↗

bioRxiv2020

MacaquePose: A novel 'in the wild' macaque monkey pose dataset for markerless motion capture

Video-based markerless motion capture permits quantification of an animals pose and motion, with a high spatiotemporal resolution in a naturalistic context, and is a powerful tool for analyzing the relationship between the animals behaviors and its brain functions. Macaque monkeys are excellent non-human primate models, especially for studying neuroscience. Due to the lack of a dataset allowing training of a deep neural network for the macaques markerless motion capture in the naturalistic context, it has been challenging to apply this technology for macaques-based studies. In this study, we created MacaquePose, a novel open dataset with manually labeled body part positions for macaques in naturalistic scenes, consisting of >13,000 images, refined by researchers. We show that the pose estimation performance of an artificial neural network trained with the dataset is close to that of a human-level. The MacaquePose will provide a platform for innovative behavior analysis for non-human primate.

neuroscience↗

bioRxiv2020

Performance Evaluation of Markerless 3D Skeleton Pose Estimates with Pop Dance Motion Sequence

The evaluation of markerless pose estimation performed by OpenPose has been getting much attention from researchers of human movement studies. This work aims to evaluate and compare the output joint positions estimated by the OpenPose with a marker-based motion-capture data recorded on a pop dance motion. Although the marker-based motion capture can accurately measure and record the human joint positions, this particular set-up is expensive. The framework to compare the outputs of the markerless method to the ground truth marker-based joint remains unknown, especially for complex body motion. Synchronization, camera calibration, and 3D reconstruction by fusing the outputs of the markerless method (OpenPose) are discussed. In this case study, the comparison results illustrate that even if the markerless method expects to fail when the subjects body parts are self-occluded, the average magnitude errors for each key points are less than 700 mm.

neuroscience↗

bioRxiv2018

Monkey Features Location Identification Using Convolutional Neural Networks

Understanding animal behavior in its natural habitat is a challenging task. One of the primary step for analyzing animal behavior is feature detection. In this study, we propose the use of deep convolutional neural network (CNN) to locate monkey features from raw RGB images of monkey in its natural environment. We train the model to identify features such as the nose and shoulders of the monkey at about 0.01 model loss.

animal behavior and cognition↗