Audio-visual scene classification: analysis of DCASE 2021 Challenge submissions
This work provides insights into effective techniques for multimodal scene classification, benefiting researchers in audio-visual AI, but it is incremental as it reviews existing methods in a challenge context.
The paper analyzed submissions to the DCASE 2021 Challenge for audio-visual scene classification, finding that top systems used pretrained models and multimodal methods, with the best achieving a logloss of 0.195 and accuracy of 93.8%, significantly outperforming the baseline.
This paper presents the details of the Audio-Visual Scene Classification task in the DCASE 2021 Challenge (Task 1 Subtask B). The task is concerned with classification using audio and video modalities, using a dataset of synchronized recordings. This task has attracted 43 submissions from 13 different teams around the world. Among all submissions, more than half of the submitted systems have better performance than the baseline. The common techniques among the top systems are the usage of large pretrained models such as ResNet or EfficientNet which are trained for the task-specific problem. Fine-tuning, transfer learning, and data augmentation techniques are also employed to boost the performance. More importantly, multi-modal methods using both audio and video are employed by all the top 5 teams. The best system among all achieved a logloss of 0.195 and accuracy of 93.8%, compared to the baseline system with logloss of 0.662 and accuracy of 77.1%.