Abstract
Gesture Recognition pertains to recognizing meaningful expressions of motion by a human, involving the hands, arms, face, head, and/or body. It is of utmost important in designing an intelligent and efficient human–computer interface. The applications of gesture recognition are manifold, ranging from sign language through medical rehabilitation, monitoring patients or elder people, surveillance systems, sports gesture analysis, human behavior analysis etc., to virtual reality. In recent years, there has been increased interest in video summarization and automatic sports highlights generation in the game of Cricket. In Cricket, the Umpire has the authority to make important decisions about events on the field. The Umpire signals important events using unique hand on signals and gestures. The primary intention of our work is to design and develop a new robust method for Umpire Action and Non-Action Gesture Identification and Recognition based on the Umpire Segmentation and the proposed Histogram Oriented Gradient (HOG) feature Extraction oriented of Deep Features. Primarily the 80% of Umpire action and non-action images in a cricket match, about 1, 93,000 frames, the Histogram of Oriented Gradient Deep Features are calculated and trained the system having six gestures of Umpire pose using Densenet121.
Description
We surveyed total of 10 Research Papers of different authors that have done projects in this field. We found that we can create dataset ourself using webcam and save in the respective directory of the sign and umpire. The images will then undergo for pre- processing techniques so that better quality of the images can be used for feature extraction. We also found about DenseNet121 Algorithm that can been implemented on the saved model for prediction. Our work will also include data gathering using a web camera to increase the dataset size to more than 50000 RGB images making our prediction more robust. Process of real-time prediction using image frames from a web camera with rates of 50 to 100 Hz is used. Currently the aim of this project is limited to detecting Umpire’s gesture. Future Enhancement may include more cameras for detecting the umpire in 360 degrees to improve gesture recognition and unique gestures for actions that do not have gestures can be added to reduce the workload on the review team. With enough funds this project can overtake and detect even the non-gesture based actions like boundary detection for run out, four and wide. Using the trained model for Gestures and Umpire detection we have got 100% accuracy. The above image shows an example of umpire detection and non- umpire detection.
Links and resources
Tags