Multilabel Genre Prediction Using Deep-Learning Frameworks

Abstract

In this study, transfer learning has been used to overcome multilabel classification tasks. As a case study, movie genre classification by using posters has been chosen. Six state-of-the-art pretrained models, VGG16, ResNet, DenseNet, Inception, MobileNet, and ConvNeXt, have been employed for this experiment. The movie posters have been obtained from Internet Movie Database (IMDB). The dataset has been divided using an iterative stratification technique. A sequence of dense layers has been added on top of each model and these models have been trained and fine-tuned. All the results of the models compared considered accuracy, loss, Hamming loss, F1-score, precision, and AUC metrics. When the metrics used were evaluated, the most successful result regarding accuracy has been obtained from the modified DenseNet architecture at 90%. Also, the ConvNeXt, which is the newest model among all, performed quite satisfactorily, reaching over 90% accuracy. This study uses an iterative stratification method to split an unbalanced dataset which provides more reliable results than the classical splitting method which is the common method in the literature. Also, the feature extraction capabilities of the six pretrained models have been compared. The outcome of this study shows promising results regarding multilabel classification. As for future work, it is planned to enhance this study by using natural language processing and ensemble methods.

Developing Smart Archives in Society 5.0: Leveraging Artificial Intelligence for Managing Audiovisual Archives in Africa

“…Machine learning plays a crucial role in managing audiovisual archives by automating tasks related to metadata generation, content recognition, enhancement, recommendation, analysis, searchability, duplication detection, preservation, and longevity. These capabilities not only streamline archival workflows but also improve the accessibility, organisation, and usability of multimedia archives for various stakeholders (Unal et al 2023).…”

“…NLP techniques can summarise textual content associated with audiovisual archives, providing concise descriptions or abstracts that capture the essence of longer documents or transcripts. Summarisation algorithms enable users to quickly grasp the main ideas and themes of archived content, aiding in content exploration and decision-making (Unal et al 2023).…”

“…NLP techniques can support question-answering systems that allow users to pose natural language queries about audiovisual content and receive relevant answers based on the content's metadata, transcripts or subtitles. Question-answering systems improve the accessibility and usability of audiovisual archives, enabling users to find specific information quickly and easily (Unal et al 2023). Natural language processing plays a vital role in audiovisual archives by enabling metadata enrichment, transcription and subtitling, semantic analysis, content summarisation, sentiment analysis, language translation, named entity recognition, and question answering.…”

“…A robotic machine can automate repetitive and time-consuming tasks involved in managing audiovisual archives, such as digitisation, cataloguing, metadata tagging, and quality control. This automation reduces the burden on human operators, increases efficiency, and accelerates the archival workflow (Unal et al 2023). Robotic machines are capable of performing tasks with high precision and accuracy, ensuring consistent quality in archival processes such as digitisation, content analysis and preservation.…”

A Natural-Language-Processing-Based Method for the Clustering and Analysis of Movie Reviews and Classification by Genre

“…According to Unal et al (2023) [8], multilabel classification involves attributing multiple labels or tags to a given input instance. Within the specific domain of movie films, this task involves recognizing and allocating appropriate genre labels to posters based on their visual attributes.…”