Simplified Training for Gesture Recognition

Faugeroux, Romain; Vieira, Thales; Martínez, Dimas; Lewiner, Thomas

doi:10.1109/sibgrapi.2014.46

Cited by 4 publications

(5 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similarly to our work, Faugeroux et al [9] also use a subset of frames extracted from the sequence. In their technique, a decision forest is built by using nodes to represent discriminative key poses and leaves to represent gestures, i.e.…”

Section: Related Workmentioning

confidence: 94%

“…In the work of Faugeroux et al [9] is presented a methodology to automatically extract a concise and discriminative key pose set as well as the key pose sequence representing each gesture example. To extract the key poses, the method initially represents a record of gestures (several gestures performed with a short pause between each one) as a curve in high dimensional space, and consider that between each gesture execution, the user performs short pauses.…”

Section: Key Poses Extractionmentioning

confidence: 99%

“…We compare the performance of our methodology against the Decision Forests adopted in [9], which also applies a filtering step to select the more discriminative …”

Section: Regularizationmentioning

confidence: 99%

“…Instead of using the traditional technique of analyzing all frames of a recording, we employ a filtering technique based on key poses [9] that selects only a small set of frames to be used by the learning method. Specifically, we employ the discriminative Latent-Dynamic Conditional Random Field (LDCRF) model [26] to capture the intrinsic and extrinsic dynamics of substructure of the moves.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

High performance moves recognition and sequence segmentation based on key poses filtering

Vicente

Nascimento

Emery

et al. 2016

2016 IEEE Winter Conference on Applications of Computer Vision (WACV)

Self Cite

View full text Add to dashboard Cite

We present a discriminative key pose-based approach for moves recognition and segmentation of training sequences for high performance sports. Compared to daily human gestures, moves in high performance sports are faster and have low inter-class variability, which produce noisy features and ambiguity. Our approach combines a robust filtering strategy to select frames composed of discriminative poses (key poses) and the discriminative Latent-Dynamic Conditional Random Fields (LDCRF) model to predict a label for each frame from the training sequence. We evaluate our approach on unsegmented sequences of Taekwondo training. Experimental results indicate that our methodology outperforms the Decision Forests method in terms of efficiency and accuracy. Our average recognition rate was equal to 74.72% while Decision Forests achieves 58.29%. The experiments also show that our approach was able to recognize and segment high speed moves like roundhouse kicks, which can reach peak linear speeds up to 26 m/s.

show abstract

Section: Related Workmentioning

confidence: 94%

Section: Key Poses Extractionmentioning

confidence: 99%

“…We compare the performance of our methodology against the Decision Forests adopted in [9], which also applies a filtering step to select the more discriminative …”

Section: Regularizationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

High performance moves recognition and sequence segmentation based on key poses filtering

Vicente

Nascimento

Emery

et al. 2016

2016 IEEE Winter Conference on Applications of Computer Vision (WACV)

Self Cite

View full text Add to dashboard Cite

show abstract

“…These applications all rely on automatic gesture recognition. This is due to the need for automated mathematical model-based semantic interpretation of human gestures [3]. Owing to the structure of vision and the way synapses are structured during the course of human brain development, body language, or gestures, can be recognized by an individual with ease [4].…”

Section: Introductionmentioning

confidence: 99%

Robust Hand Gesture Recognition Using HOG Features and machine learning

Sayed,

Bakheet,

Mofaddel

et al. 2024

Sohag Journal of Sciences

View full text Add to dashboard Cite

Physical disability is one aspect of people that cannot be disregarded. A deaf person is someone who is naturally unable to hear.A unique language known as "Sign-Language" is used to represent their expertise. Sign language is one of the most popular forms of deaf people to learn is American Sign Language (ASL).A collection of images of hands in various hand gestures or shapes are used in American Sign Language. In this study, we introduce feature-based algorithmic analysis to create a significant model for American Sign Language hand gesture identification. This model can be used to effectively learn in order to make a machine intelligent. We create a list of helpful features from digital images of hand gestures for efficient machine learning.For the preprocessing process, a histogram equalization technique and the an-isotropic diffusion filter are used. To extract image features a robust histogram of oriented gradient feature extraction method is proposed then three different machine learning classifiers are performed to achieve the classification process. To test our model, experiments are achieved using the American MNIST sign language dataset. With the use of HOG as feature and Support Vector Machine as classifier, the system yields by achieving high levels of sensitivity, specificity, and accuracy (99.8%, 98.9% and 99.6%, respectively). We can derive that the proposed model is an efficient sign language detection system. keywords:computer vision, machine learning,histogram oriented gradient , hand-gesture recognition.

show abstract