Where is my hand? Deep hand segmentation for visual self-recognition in humanoid robots
نویسندگان
چکیده
The ability to distinguish between the self and background is of paramount importance for robotic tasks. particular case hands, as end effectors a system that more often enter into contact with other elements environment, must be perceived tracked precision execute intended tasks dexterity without colliding obstacles. They are fundamental several applications, from Human–Robot Interaction object manipulation. Modern humanoid robots characterized by high number degrees freedom which makes their forward kinematics models very sensitive uncertainty. Thus, resorting vision sensing can only solution endow these good perception self, being able localize body parts precision. In this paper, we propose use Convolution Neural Network (CNN) segment robot hand an image in egocentric view. It known CNNs require huge amount data trained. To overcome challenge labeling real-world images, simulated datasets exploiting domain randomization techniques. We fine-tuned Mask-RCNN network specific task segmenting Vizzy. focus our attention on developing methodology requires low amounts achieve reasonable performance while giving detailed insight how properly generate variability training dataset. Moreover, analyze fine-tuning process within complex model Mask-RCNN, understanding weights should transferred new hands. Our final was trained solely synthetic images achieves average IoU 82% validation 56.3% real test data. These results were achieved 1000 3 h time using single GPU.
منابع مشابه
Segmentation Methods for Hand Written Character Recognition
Hand written Character Recognition is area of research since many years. Automation of existing manual system is need of most industries as well as government areas. Recognition of hand written characters is a demand for many fields. In this paper we have discussed our approach for hand written character segmentation. This paper discusses various methodologies to segment a text based image at v...
متن کاملReal Time Hand Gesture Recognition Including Hand Segmentation and Tracking
In this paper we present a system that performs automatic gesture recognition. The system consists of two main components: (i) A unified technique for segmentation and tracking of face and hands using a skin detection algorithm along with handling occlusion between skin objects to keep track of the status of the occluded parts. This is realized by combining 3 useful features, namely, color, mot...
متن کاملVisual Recognition of Hand Motion
...................................................................................................................... iii Preface........................................................................................................................... v Acknowledgments................................................................................................... vii
متن کامل3D Hand Tracker for Visual Sign Recognition
The gesture recognition process, in general, may be divided into two stages: motion sensing, which extracts useful data from hand motion; and the classification process, which classifies the motion sensing data as gestures. We have developed the visionbased Hand Motion Understanding (HMU) system that recognises static and dynamic Australian Sign Language (Auslan) signs by extracting and classif...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Robotics and Autonomous Systems
سال: 2021
ISSN: ['0921-8890', '1872-793X']
DOI: https://doi.org/10.1016/j.robot.2021.103857