Prediction, Random Forest, Neural Networks, LeNet, AlexNet, VGG-16, VGG-19, ResNet, Logistic Regression, TensorFlow, Adam Optimizer, SGD Optimizer, Keras, CNN, Image Classification, Machine Learning, Deep Learning
This paper proposes a Convolutional Neural Network (CNN) based method for automatic movie genre classification from poster images. The proposed CNN architecture utilizes a multi-layered structure trained on a large dataset of movie posters. The paper details the CNN design, training process, and data pre-processing techniques employed. Data pre-processing includes one-hot encoding genre labels, handling missing values, addressing data imbalance, and image re-sizing/normalization. The performance of the proposed CNN is evaluated and compared against established models like LeNet, AlexNet, VGG variants, ResNet-50, Logistic Regression, and Random Forest.
Manually categorizing movies based on their genre from posters is inefficient, especially with the ever-growing number of films released annually. Automated systems can help streamline this process, offering quick, accurate, and scalable solutions for large movie databases. This paper investigates the use of CNNs for automatic movie genre classification from poster images, aiming to build an effective and efficient model capable of performing genre classification in real-world applications.
In order to run the above files, download the necessary libraries using the command, " pip install [Libraryname] " or if running in Jupyter Notebooks then use command " !pip install [Libraryname] ". These codes might take huge amount of time to train as we have perfoemed several experimental analysis and also it can vary based on the computing powers.