|
|
6-DOF Object Localization in Industrial Applications
Macurová, Nela ; Španěl, Michal (referee) ; Hradiš, Michal (advisor)
The aim of this work is to design a method for the object localization in the point could and as accurately as possible estimates the 6D pose of known objects in the industrial scene for bin picking. The design of the solution is inspired by the PoseCNN network. The solution also includes a scene simulator that generates artificial data. The simulator is used to generate a training data set containing 2 objects for training a convolutional neural network. The network is tested on annotated real scenes and achieves low success, only 23.8 % and 31.6 % success for estimating translation and rotation for one type of obejct and for another 12.4 % and 21.6 %, while the tolerance for correct estimation is 5 mm and 15°. However, by using the ICP algorithm on the estimated results, the success of the translation estimate is 81.5 % and the rotation is 51.8 % and for the second object 51.9 % and 48.7 %. The benefit of this work is the creation of a generator and testing the functionality of the network on small objects
|
|
|
Self-Supervised Learning for Recognition of Sports Poses in Image
Olekšák, Samuel ; Kocur, Viktor (referee) ; Herout, Adam (advisor)
This thesis demonstrates a solution for minimizing the amount of necessary labelled training data in the classification of sports poses using a neural network trained with contrastive self-supervised learning. Training consists of two stages. The first stage trains a feature extractor which uses unlabelled training images extracted from recordings of exercises from multiple viewpoints. In the second stage, using a small amount of labelled data, a simple classifier connected to the feature extractor is trained. The thesis discusses classification in the context of yoga poses, however, the final solution can be easily applied to any other sport in case of obtaining a suitable dataset. During the development of the solution, emphasis is placed on the performance of the resulting model so that it can be used on mobile devices. The resulting model reached an accuracy of 76 % using augmentations with a data set containing four labelled images per yoga pose. On a larger data set with 800 labelled images for all poses, an accuracy of 82 % is reached.
|
|
|
Understanding of Badminton Videos
Mašláň, Vojtěch ; Hradiš, Michal (referee) ; Herout, Adam (advisor)
The aim of this thesis was to experiment with machine learning models for the understanding of badminton videos. The thesis maps the current state of using computer vision and machine learning for sport videos analysis. In the experimental part of the work, several models for badminton stroke detection were made. All the models are predicting the strokes based on player poses extracted from a pose estimation model. Developed models achieved an accuracy of 80.1 % for detecting 7 different strokes and 84.0 % for detecting 4 different strokes. Trained models were then used to create a simple web application for short badminton video analysis.
|
|
|
Self-Supervised Learning for Recognition of Sports Poses in Image
Olekšák, Samuel ; Kocur, Viktor (referee) ; Herout, Adam (advisor)
This thesis demonstrates a solution for minimizing the amount of necessary labelled training data in the classification of sports poses using a neural network trained with contrastive self-supervised learning. Training consists of two stages. The first stage trains a feature extractor which uses unlabelled training images extracted from recordings of exercises from multiple viewpoints. In the second stage, using a small amount of labelled data, a simple classifier connected to the feature extractor is trained. The thesis discusses classification in the context of yoga poses, however, the final solution can be easily applied to any other sport in case of obtaining a suitable dataset. During the development of the solution, emphasis is placed on the performance of the resulting model so that it can be used on mobile devices. The resulting model reached an accuracy of 76 % using augmentations with a data set containing four labelled images per yoga pose. On a larger data set with 800 labelled images for all poses, an accuracy of 82 % is reached.
|
|
|
6-DOF Object Localization in Industrial Applications
Macurová, Nela ; Španěl, Michal (referee) ; Hradiš, Michal (advisor)
The aim of this work is to design a method for the object localization in the point could and as accurately as possible estimates the 6D pose of known objects in the industrial scene for bin picking. The design of the solution is inspired by the PoseCNN network. The solution also includes a scene simulator that generates artificial data. The simulator is used to generate a training data set containing 2 objects for training a convolutional neural network. The network is tested on annotated real scenes and achieves low success, only 23.8 % and 31.6 % success for estimating translation and rotation for one type of obejct and for another 12.4 % and 21.6 %, while the tolerance for correct estimation is 5 mm and 15°. However, by using the ICP algorithm on the estimated results, the success of the translation estimate is 81.5 % and the rotation is 51.8 % and for the second object 51.9 % and 48.7 %. The benefit of this work is the creation of a generator and testing the functionality of the network on small objects
|
guest ::
