The use of computers to read musical scores is referred to as optical music recognition (OMR). The recent advancements in artificial intelligence and big data have led to the development of deep learning approaches for recognizing musical notes. Previous research has shown that there is a lot of room for improvement in handwritten musical notation recognition systems due to differences in writing styles and the complex structure of musical symbols. The research described here aims to develop a deep learning-based system for recognizing handwritten musical notation. The system uses a convolutional neural network (CNN) to extract and learn pixel features of musical symbols and achieve a recognition accuracy of over 90%. The CNN model was trained using image samples from the HOMUS dataset and fine-tuned to minimize the loss function and reduce classification errors. The CNN model achieved an accuracy of 96.95% on the test samples, which is a significant improvement over the 86.0% accuracy from previous studies. The performance of the CNN model was also compared to five state-of-the-art deep learning methods, namely, quantum gray Wolf optimization (QGWO) algorithm, nonfully connected network (NFC-Net) classifier, nearest neighbor classifier, data augmentation and ensemble learning, and the CNN model outperformed four of them. However, the CNN model occasionally misclassified musical symbols with similar shapes, indicating that there is still room for improvement in the system’s performance. Future research could focus on improving the model’s performance on similar-shaped symbols. Overall, the research demonstrates the effectiveness of using a CNN model for handwritten musical notation recognition and highlights the potential of deep learning approaches in this area.

Recent studies show that introducing phonetic information into multi-task learning could significantly improve the performance of speaker embedding extraction. However, benefits of such architectures usually depend largely on the availibility of a well-matched dataset, and domain or language mismatch would result in obvious dropdown in performance. Meanwhile, the utilization of these massive mismatched data and application of these auxiliary tasks may bring many rich features that could be exploited. In this paper, we propose a phoneme-aware adaptation network with discrepancy minimization and dynamically-classified vector for text-independent speaker verification to address these abovementioned challenges. More specifically, our method first utilize the maximum mean discrepancy (MMD) as part of the total loss function to solve the mismatch between training data of the speaker subnet and the phoneme subnet. And then we use a dynamically-classified vector-guided softmax loss (DV-Softmax), which could adaptively emphasize different high-quality features and dynamically change their weights, to guide the discriminative speaker embedding. Experimental results on VoxCeleb1 data set confirmed its superiority against the other state-of-the-art phoneme adaptation methods, providing approximately 15% relative improvements in equal error rate (EER).