Filters








5,347 Hits in 4.9 sec

Joint Visual-Temporal Embedding for Unsupervised Learning of Actions in Untrimmed Sequences [article]

Rosaura G. VidalMata, Walter J. Scheirer, Anna Kukleva, David Cox, Hilde Kuehne
2020 arXiv   pre-print
To address this problem, this paper proposes an approach for the unsupervised learning of actions in untrimmed video sequences based on a joint visual-temporal embedding space.  ...  Understanding the structure of complex activities in untrimmed videos is a challenging task in the area of action recognition.  ...  an embedding space for unsupervised actions recognition.  ... 
arXiv:2001.11122v3 fatcat:serd5ckhjfhftheykadzyay72u

Unsupervised Learning and Segmentation of Complex Activities from Video [article]

Fadime Sener, Angela Yao
2018 arXiv   pre-print
This paper presents a new method for unsupervised segmentation of complex activities from video into multiple steps, or sub-activities, without any textual input.  ...  We propose an iterative discriminative-generative approach which alternates between discriminatively learning the appearance of sub-activities from the videos' visual features to sub-activity labels and  ...  Our objective in learning the embeddings is to cluster the video features discriminatively.  ... 
arXiv:1803.09490v1 fatcat:6tjjjylidbgqximxelrmztxyxy

Unsupervised Learning and Segmentation of Complex Activities from Video

Fadime Sener, Angela Yao
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
This paper presents a new method for unsupervised segmentation of complex activities from video into multiple steps, or sub-activities, without any textual input.  ...  We propose an iterative discriminative-generative approach which alternates between discriminatively learning the appearance of sub-activities from the videos' visual features to sub-activity labels and  ...  Our objective in learning the embeddings is to cluster the video features discriminatively.  ... 
doi:10.1109/cvpr.2018.00873 dblp:conf/cvpr/SenerY18 fatcat:kqefuecyzbbzdg3gkmvxh3ps44

SSCAP: Self-supervised Co-occurrence Action Parsing for Unsupervised Temporal Action Segmentation [article]

Zhe Wang, Hao Chen, Xinyu Li, Chunhui Liu, Yuanjun Xiong, Joseph Tighe, Charless Fowlkes
2021 arXiv   pre-print
the structure of activities, but also estimate the temporal path of the sub-actions in an accurate and general way.  ...  We evaluate on both classic datasets (Breakfast, 50Salads) and the emerging fine-grained action dataset (FineGym) with more complex activity structures and similar sub-actions.  ...  Acknowledgement We thank reviewers for the valuable suggestions. Zhe Wang personally thanks Fanyi Xiao, Bing Shuai and Uta Büchler for helpful discussion  ... 
arXiv:2105.14158v3 fatcat:qzq4isslj5hkpgjl7tcfcbxhg4

Motion2Vec: Semi-Supervised Representation Learning from Surgical Videos [article]

Ajay Kumar Tanwani, Pierre Sermanet, Andy Yan, Raghav Anand, Mariano Phielipp, Ken Goldberg
2020 arXiv   pre-print
In this paper, we learn a motion-centric representation of surgical video demonstrations by grouping them into action segments/sub-goals/options in a semi-supervised manner.  ...  Learning meaningful visual representations in an embedding space can facilitate generalization in downstream tasks such as action segmentation and imitation.  ...  Learning from multiple viewpoints, temporal sequences, labeled action segments, weakly supervised signals such as order of sub-actions, text-based annotations or unsupervised learning are feasible alternatives  ... 
arXiv:2006.00545v1 fatcat:l7r5yhmm5jbmtckxrhi43xacuu

Unsupervised learning of action classes with continuous temporal embedding [article]

Anna Kukleva, Hilde Kuehne, Fadime Sener, Juergen Gall
2019 arXiv   pre-print
To address this issue, we propose an unsupervised approach for learning action classes from untrimmed video sequences.  ...  One problem in this context arises from the need to define and label action boundaries to create annotations for training which is very time and cost intensive.  ...  They introduced an iterative approach which alternates between discriminative learning of the appearance of sub-activities from visual features and generative modeling of the temporal structure of sub-activities  ... 
arXiv:1904.04189v1 fatcat:3v4wa3manzetnjl4df2bqltsjq

Unsupervised Feature Learning of Human Actions as Trajectories in Pose Embedding Manifold [article]

Jogendra Nath Kundu, Maharshi Gor, Phani Krishna Uppala, R. Venkatesh Babu
2018 arXiv   pre-print
We demonstrate state-of-the-art transfer-ability of the learned representation against other supervisedly and unsupervisedly learned motion embeddings for the task of fine-grained action recognition on  ...  In contrast to end-to-end framework explored by previous works, we disentangle the task of individual pose representation learning from the task of learning actions as a trajectory in pose embedding space  ...  Acknowledgements This work was supported by a CSIR Fellowship (Jogendra), and a project grant from Robert Bosch Centre for Cyber-Physical Systems, IISc.  ... 
arXiv:1812.02592v1 fatcat:37jnz4444faudnvaiao5d6sjym

Combining Supervised and Unsupervised Learning Algorithms for Human Activity Recognition

Elena-Alexandra Budisteanu, Irina Georgiana Mocanu
2021 Sensors  
The method combines supervised and unsupervised learning algorithms in order to provide qualitative results and performance in real time.  ...  Recent methods employ supervised and unsupervised deep learning techniques in which spatial and temporal dependency is modeled.  ...  The first part, the Discriminant model, describes an unsupervised manner of grouping up activities based on a similarity measure, using K-Means or GMM over an embedding space of an autoencoder.  ... 
doi:10.3390/s21186309 pmid:34577515 pmcid:PMC8473063 fatcat:lbueq4nijzdsldeyjzmhpwcmbu

Self-Discriminative Learning for Unsupervised Document Embedding

Hong-You Chen, Chin-Hua Hu, Leila Wehbe, Shou-De Lin
2019 Proceedings of the 2019 Conference of the North  
Unsupervised document representation learning is an important task providing pre-trained features for NLP applications.  ...  embedding space with a discriminative network and a novel objective.  ...  Algorithm 1 Self-Discriminative Learning for Unsupervised Document Embedding Input: Documents X = {X i } n 1 , p, k, h w , h s .  ... 
doi:10.18653/v1/n19-1255 dblp:conf/naacl/ChenHWL19 fatcat:kcs6yhybjja3vmun6325npeyjm

Unsupervised Video Understanding by Reconciliation of Posture Similarities [article]

Timo Milbich, Miguel Bautista, Ekaterina Sutter, Bjorn Ommer
2017 arXiv   pre-print
Understanding human activity and being able to explain it in detail surpasses mere action classification by far in both complexity and value.  ...  a single concerted pose embedding despite changes in appearance across sequences.  ...  Acknowledgements: This work has been supported in part by the Heidelberg Academy for the Sciences. We are grateful to the NVIDIA corporation for donating a Titan X GPU.  ... 
arXiv:1708.01191v1 fatcat:uczfnfnp2fhnjki7bnpph3jttu

Fine-grained Action Segmentation using the Semi-Supervised Action GAN [article]

Harshala Gammulle, Simon Denman, Sridha Sridharan, Clinton Fookes
2019 arXiv   pre-print
The GAN is made to learn features in a semi-supervised manner, enabling the model to perform action classification jointly with the standard, unsupervised, GAN learning procedure.  ...  The challenge for this task lies in the need to represent the hierarchical nature of the actions and to detect the transitions between actions, allowing us to localise the actions within the video effectively  ...  For different peaks and valleys in the activations, we show the embedding that has been stored at the time step.  ... 
arXiv:1909.09269v1 fatcat:ys4oklz3yfgxlb2xcaj7j7hf3m

Unsupervised Action Segmentation by Joint Representation Learning and Online Clustering [article]

Sateesh Kumar, Sanjay Haresh, Awais Ahmed, Andrey Konin, M. Zeeshan Zia, Quoc-Huy Tran
2022 arXiv   pre-print
The temporal optimal transport module enables our approach to learn effective representations for unsupervised activity segmentation.  ...  We present a novel approach for unsupervised activity segmentation which uses video frame clustering as a pretext task and simultaneously performs representation learning and online clustering.  ...  one of the action/sub-activity classes.  ... 
arXiv:2105.13353v5 fatcat:y4kafit2i5cfzpfxmkjvinvezq

A Perceptual Prediction Framework for Self Supervised Event Segmentation [article]

Sathyanarayanan N. Aakur, Sudeep Sarkar
2019 arXiv   pre-print
We also show that the proposed approach is able to learn highly discriminative features that help improve action recognition when used in a representation learning paradigm.  ...  In this paper, we tackle the problem of self-supervised temporal segmentation of long videos that alleviate the need for any supervision.  ...  There also exists a "background activities" which consists of sequence where there does not exist a clear sub-activity that is visually discriminable.  ... 
arXiv:1811.04869v3 fatcat:suoguja4dzbvrjsg3m5lkcdtnq

Learning object, grasping and manipulation activities using hierarchical HMMs

Mitesh Patel, Jaime Valls Miro, Danica Kragic, Carl Henrik Ek, Gamini Dissanayake
2014 Autonomous Robots  
This article presents a probabilistic algorithm for representing and learning complex manipulation activities performed by humans in everyday life.  ...  Activity data from 3D video sequencing of human manipulation of different objects handled in everyday life is used for evaluation.  ...  their contribution towards acquiring the data sets used in this paper.  ... 
doi:10.1007/s10514-014-9392-1 fatcat:c5c2hsf3lvhbbnb7rz4aarihla

An Unsupervised Framework for Online Spatiotemporal Detection of Activities of Daily Living by Hierarchical Activity Models

Farhood Negin, François Brémond
2019 Sensors  
Moreover, the local dynamic information incorporates complex local motion patterns of daily activities into the models.  ...  The experimental data on a variety of monitoring scenarios in hospital settings reveals how this framework can be exploited to provide timely diagnose and medical interventions for cognitive disorders,  ...  We use attributes of an activity and its sub-activities for modeling and accordingly, learning is performed automatically using the DAs and PEs in different resolutions.  ... 
doi:10.3390/s19194237 pmid:31569564 pmcid:PMC6806106 fatcat:7ivaczgntbhhzfl6ylyau2epmm
« Previous Showing results 1 — 15 out of 5,347 results