Cutting-edge visual computing techniques, especially those recognized for their large-scale capabilities, have exhibited impressive achievements in computer vision area. However, the significant reliance on vast amounts of data, label, and computational resources of these visual computing techniques present challenges when deploying to real-world scenarios with limited resources. This workshop aims to bring together experts in data-efficient, label-efficient, and computation-efficient visual computing fields, establishing a collaborative platform to exchange recent breakthroughs and deliberate on the future direction of visual computing models. By facilitating the exchange of ideas and insights, we aspire to address the efficiency challenges inherent in visual computing and contribute to the evolution of its practical applications in the real world.

This workshop centers on the academic exploration of efficient methodologies within the realm of visual computing. Our focus includes data-efficient strategies like image/video-compression, label- efficient strategies like zero/few-shot learning, and model-efficient approaches like model sparsification/quantization. By convening researchers specializing in these areas, we aim to facilitate the sharing of recent research findings and engage in discussions about the future trajectories of efficient visual computation. The importance of this workshop is emphasized by the current relevance of its theme, which has received considerable attention from researchers because of its direct relevance to practical uses. Through this platform, we aim to facilitate the sharing of ideas, creating a space for the expression of innovative perspectives aimed at tackling the efficiency issues faced by visual computing. This endeavor aims to contribute to the progress of real-world applications.