The cocktail party problem can be more effectively addressed by leveraging the speaker’s visual and audio information. This paper proposes a method to improve the audio’s separation using two visual cues: facial features and lip movement. Firstly, residual connections are introduced in the audio separation module to extract detailed features. Secondly, considering the video stream contains information other than the face, which has a minimal correlation with the audio, an attention mechanism is employed in the face module to focus on crucial information. Then, the loss function considers the audio-visual similarity to take advantage of the relationship between audio and visual completely. Experimental results on the public VoxCeleb2 dataset show that the proposed model significantly enhanced SDR, PSEQ, and STOI, especially 4 dB improvements in SDR.

Modeling and simulating the underwater optical imaging process can assist in optimizing the configuration of underwater optical imaging technology. Based on the Monte Carlo (MC) method, we propose an optical imaging model which is tailored for deep-sea luminescent objects. Employing GPU parallel acceleration expedites the speed of MC simulation and ray-tracing, achieving a three-order-of-magnitude speedup over a CPU-based program. A deep-sea single-lens imaging system is constructed in the model, composed of a luminescent object, water medium, double-convex lens, aperture diaphragm, and sensor. The image of the luminescent object passing through the imaging system is generated using the forward ray-tracing method. This model enables an intuitive analysis of the inherent optical properties of water and imaging device parameters, such as sensor size, lens focal length, field of view (FOV), and camera position on imaging outcomes in the deep-sea environment.