Activity detection and recognition of daily living events

Avgerinakis, Konstantinos; Briassouli, Alexia; Kompatsiaris, Ioannis

doi:10.1145/2505323.2505327

Cited by 13 publications

(6 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Clustering is one of the most used techniques in the recognition of Activities of Daily Living [2][3][4][5][6][7][8][9][10][11]. Clustering analysis is the formal study of algorithms and methods to group objects according to measurements, perceived attributes, intrinsic features or likelihoods [12].…”

Section: Clusteringmentioning

confidence: 99%

“…Therefore, it is necessary to get information about ADL performed by the person in a real-life scenario. For upperlimb movements, the most common sensors are accelerometers and gyroscopes [2][3][4][5]. However, the recognition of activities by using acceleration signals or any other device or combination of them have not shown acceptable results in an extensive set of activities.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Clustering Techniques for Recognition of ADL Performed with Upper Limbs

Parra-Sánchez¹,

Gómez-González²,

Quintero-Ortega³

et al. 2019

RCS

View full text Add to dashboard Cite

Activities of Daily Living (ADL) have a background of selfsufficiency and survival function. Upper limbs participate actively in many ADL; particularly, activities related to feeding, communication, and grooming. The performance of such activities is a parameter of independence. Various researchers have studied ADL in a free-living environment by using inertial sensors. However, functional-activity recognition with low recognition rate is a persistent result. This work proposes the use of well-known clustering techniques for ADL recognition by using as feeding signal the vertical trajectory of wrist relative to the shoulder.

show abstract

Section: Clusteringmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Clustering Techniques for Recognition of ADL Performed with Upper Limbs

Parra-Sánchez¹,

Gómez-González²,

Quintero-Ortega³

et al. 2019

RCS

View full text Add to dashboard Cite

show abstract

“…For daily-living activities, less methods and datasets for detection were introduced. In [17] the authors used a simple method for detection depending on the person's motion. They segment chunks for successive frames that contain motion and keep track of all interest points, then pass it to action recognition stage where HOG-HOF features around the sampled interest points are computed followed by Bag of Words representation and SVM classifier.…”

Section: Related Workmentioning

confidence: 99%

Online temporal detection of daily-living human activities in long untrimmed video streams

Goel

Abubakr

Koperski

et al. 2018

2018 IEEE International Conference on Image Processing, Applications and Systems (IPAS)

View full text Add to dashboard Cite

Many approaches were proposed to solve the problem of activity recognition in short clipped videos, which achieved impressive results with hand-crafted and deep features. However, it is not practical to have clipped videos in real life, where cameras provide continuous video streams in applications such as robotics, video surveillance, and smart-homes. Here comes the importance of activity detection to help recognizing and localizing each activity happening in long videos. Activity detection can be defined as the ability to localize starting and ending of each human activity happening in the video, in addition to recognizing each activity label. A more challenging category of human activities is the daily-living activities, such as eating, reading, cooking, etc, which have low inter-class variation and environment where actions are performed are similar. In this work we focus on solving the problem of detection of daily-living activities in untrimmed video streams. We introduce new online activity detection pipeline that utilizes single sliding window approach in a novel way; the classifier is trained with sub-parts of training activities, and an online frame-level early detection is done for sub-parts of long activities during detection. Finally, a greedy Markov model based post processing algorithm is applied to remove false detection and achieve better results. We test our approaches on two daily-living datasets, DAHLIA and GAADRD, outperforming state of the art results by more than 10%.

show abstract

“…In [50], the authors combined this approach with a similar feature for oriented optical flow (histogram of oriented flow -HOF-) in order to capture both shape and motion, leading to the state of the art standard for human detection. In [51], this method is used in addition to holistic features extracted from raw trajectory cues for recognition of ADL of healthy subjects and people with dementia using the URADL dataset [52].…”

Section: Feature Extraction and Recognitionmentioning

confidence: 99%