This paper is published in Volume-6, Issue-3, 2020
Area
Information Science
Author
Puja Kedia, Saurabh Singh, Lakshmi Saai Rasazna Konagalla, Shantha H. Biradar
Org/Univ
Sir M. Visvesvaraya Institute of Technology, Bengaluru, Karnataka, India
Pub. Date
05 June, 2020
Paper ID
V6I3-1425
Publisher
Keywords
CNN, Drop-Out Rate, Keras, Activation Function, Overfitting, Cross-Validation

Citationsacebook

IEEE
Puja Kedia, Saurabh Singh, Lakshmi Saai Rasazna Konagalla, Shantha H. Biradar. Duration of an actor in a video using Keras, International Journal of Advance Research, Ideas and Innovations in Technology, www.IJARIIT.com.

APA
Puja Kedia, Saurabh Singh, Lakshmi Saai Rasazna Konagalla, Shantha H. Biradar (2020). Duration of an actor in a video using Keras. International Journal of Advance Research, Ideas and Innovations in Technology, 6(3) www.IJARIIT.com.

MLA
Puja Kedia, Saurabh Singh, Lakshmi Saai Rasazna Konagalla, Shantha H. Biradar. "Duration of an actor in a video using Keras." International Journal of Advance Research, Ideas and Innovations in Technology 6.3 (2020). www.IJARIIT.com.

Abstract

This research is mainly based on automating the process of calculating the time taken for any actor to appear on the screen. It is one of the main factors for determining the remuneration to be given to actors for appearing on the screen for a particular time period. By automating this process, we can accurately determine the screen time of an actor with minimum error. This proposed idea can be implemented using the knowledge related to image processing with the help of CNN architecture. The major part of the research lies in determining the hyperparameters and the right model that fits the given video appropriately and gives the best results for the model evaluation. The major findings of this paper lie on analyzing the right activation function, the number of layers for the neural network, finding the drop-out rate for the trained neural network, deciding upon the weight sharing of the input attributes, and of the hidden layers. The final outcome of this performed experiment is a neural network that can be used for deciding the duration of an actor on the screen.