Discovering Cricket Stroke Classes in Trimmed Telecast Videos

Gupta, Arpan; Karel, Ashish; Balan, M. Sakthi

doi:10.1007/978-981-15-4018-9_45

Cited by 7 publications

(8 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this work, we use the dataset of Gupta et al [5] with 562 stroke labels in 26 untrimmed Highlights videos of T20 World Cup 2016. They have dealt with the problem of temporal cricket stroke localization in [6,7].…”

Section: Literature Surveymentioning

confidence: 99%

“…They have dealt with the problem of temporal cricket stroke localization in [6,7]. Gupta et al [5] have used an unsupervised clustering approach on the averaged motion features of the strokes and visualized them for the same dataset.…”

Section: Literature Surveymentioning

confidence: 99%

“…We extracted Farneback [3] dense optical flow based features for all the trimmed strokes. The generated flow fields constitute the magnitude and angle maps, calculated for all consecutive frame pairs (details in [5]). The magnitude denotes the amount of displacement of a pixel while the angle denotes the direction of displacement represented in 2D plane.…”

Section: Feature Extractionmentioning

confidence: 99%

“…They capture a general trend of motion in the sequence of frames by finding a histogram over the bin values. We use an unweighted variant of the traditional HOOF feature (similar to [5]) by choosing the orientations (Θ) for only those pixels which have a corresponding magnitude (M ) above a threshold (m th ), as shown in Eq. 1.…”

Section: Orientation Histogram (Hoof)mentioning

confidence: 99%

“…While training, the temporal action labels were used and at validation/test time the evaluation was carried out similar to [4]. The optical flow based features and ResNet extracted feature variants were tested on the Cricket strokes dataset [5] having 562 trimmed strokes, with 351 in training set, 105 in validation set and 106 in the test set, labeled with 3 classes Left, Right and Ambiguous based on the direction of camera motion. We independently labeled all the strokes at a finer level, into 5 categories, based on the direction of stroke hit, as shown in Fig.…”

Section: Experimentationmentioning

confidence: 99%

See 4 more Smart Citations

Cricket Stroke Recognition Using Hard and Soft Assignment Based Bag of Visual Words

Gupta

Karel

Muthiah

2021

Communications in Computer and Information Science

Self Cite

View full text Add to dashboard Cite

In this work, we deal with the problem of activity recognition in Cricket telecast videos. We present a supervised approach for recognizing Cricket stroke categories using two variants of Bag of Visual words (BoV) model applied on dense optical flow based motion feature i.e., grid-based flattened vectors and orientation histograms, 3D ResNet extracted features and 2D ResNet extracted spatial features. These globally extracted features, in spite of the noise due to camera motion, give good results on the Cricket strokes dataset having 562 trimmed stroke videos. We independently labeled the strokes based on the direction of stroke play and the direction of camera motion, into five and three categories. respectively, and provide experimental analysis on Hard Assignment (HA) and Soft Assignment (SA) based BoV methods.Our contribution lies in showing the effectiveness of the unordered BoV representation for direction based action recognition. The experimental analysis presented in our work provides an insight into the BoV HA and SA models applied to global frame-level feature descriptors. Moreover, our experiments suggest the orientation histograms to be simple yet effective for direction based recognition tasks. The best accuracy achieved on the Cricket strokes test partition was 85.85% with orientation histograms and 82.08% with grid features, for 3 category and 5 category labels, respectively.

show abstract

Section: Literature Surveymentioning

confidence: 99%