Journal of Signal Processing Systems 2024

3 papers
The 2024 edition of the Journal of Signal Processing Systems featured research focused on the application of deep learning and signal processing techniques across computer vision, natural language processing, and model security. Key themes included the development of multimodal fusion models for action recognition and the advancement of interpretability methods for convolutional neural networks (CNNs). Specifically, research addressed the security of image-based big data models through the SIS-CAM analysis method and explored controlled text generation tailored for social network environments. Contributions from authors such as Haifeng Sang, Fang Xu, and Na Yang highlight a trend toward enhancing model transparency and cross-modal data integration. For instance, the "Supernode Fusion Model" investigated bimodal action recognition, while other studies focused on topic-oriented text generation. These works reflect an emphasis on refining neural architectures for specialized data environments and ensuring the reliability of large-scale image processing systems.

Papers

Title
SIS-CAM: An Interpretability Analysis Method for the Security of Convolutional Neural Network Models Based on Image Big Data
Supernode Fusion Model Based on Bimodal Action Recognition
Topic-Oriented Controlled Text Generation for Social Networks