NDERWATER acoustic target recognition (UATR) is crucial for applications such as environmental monitoring, Deep learning models, such as convolutional neural networks exploration, and ship noise characterization, aiding in (CNNs), excel in feature representation and transfer learning, marine resource management and ocean-based technologies to adapting well to underwater acoustics when pre-trained on enhance ocean monitoring [1], [2]. Passive sonar uses external large vision datasets [11]-[13]. Similarly, pre-trained audio acoustic signals to identify underwater objects without emitting neural networks (PANNs) [14], trained on a large audio dataset sound [2]. Spectrograms, generated through signal processing (AudioSet [15]), have proven effective for passive sonar techniques like Short-Time Fourier Transform (STFT) classification where data scarcity is a challenge [16]. Moreover, and Mel-frequency spectrograms, transform signals into visual transformer-based models, including vision transformers representations, facilitating complex pattern extraction from (ViTs) [17] and audio spectrogram transformers (ASTs) [18], acoustic data [3]-[5].

Transfer learning is commonly employed to leverage large, pre-trained models and perform fine-tuning for downstream tasks. The most prevalent pre-trained models are initially trained using ImageNet. However, their ability to generalize can vary across different data modalities. This study compares pre-trained Audio Neural Networks (PANNs) and ImageNet pre-trained models within the context of underwater acoustic target recognition (UATR). It was observed that the ImageNet pre-trained models slightly out-perform pre-trained audio models in passive sonar classification. We also analyzed the impact of audio sampling rates for model pre-training and fine-tuning. This study contributes to transfer learning applications of UATR, illustrating the potential of pre-trained models to address limitations caused by scarce, labeled data in the UATR domain.

While deep learning has reduced the prevalence of manual feature extraction, transformation of data via feature engineering remains essential for improving model performance, particularly for underwater acoustic signals. The methods by which audio signals are converted into time-frequency representations and the subsequent handling of these spectrograms can significantly impact performance. This work demonstrates the performance impact of using different combinations of time-frequency features in a histogram layer time delay neural network. An optimal set of features is identified with results indicating that specific feature combinations outperform single data features.

Underwater acoustic target detection in remote marine sensing operations is challenging due to complex sound wave propagation. Despite the availability of reliable sonar systems, target recognition remains a difficult problem. Various methods address improved target recognition. However, most struggle to disentangle the high-dimensional, non-linear patterns in the observed target recordings. In this work, a novel method combines a time delay neural network and histogram layer to incorporate statistical contexts for improved feature learning and underwater acoustic target classification. The proposed method outperforms the baseline model, demonstrating the utility in incorporating statistical contexts for passive sonar target recognition. The code for this work is publicly available.