The study introduces PROST, a framework for quantitative recognition of spatial transcriptomic patterns, consisting of two main components: the PROST Index (PI) and the PROST Neural Network (PNN). PI characterizes spatial variations in gene expression patterns through a novel index, while PNN uses a self-attention mechanism to cluster spatial domains. PROST demonstrates superior performance in identifying spatially variable genes (SVGs) and segmenting spatial domains across various spatial resolutions, from multicellular to cellular levels. The PROST Index can prioritize spatial expression variations, facilitating biological insights. The framework is validated on datasets from different platforms and resolutions, showing robustness and adaptability. PROST's ability to integrate spatial information and gene expression profiles makes it a flexible and powerful tool for analyzing spatial transcriptomic data.The study introduces PROST, a framework for quantitative recognition of spatial transcriptomic patterns, consisting of two main components: the PROST Index (PI) and the PROST Neural Network (PNN). PI characterizes spatial variations in gene expression patterns through a novel index, while PNN uses a self-attention mechanism to cluster spatial domains. PROST demonstrates superior performance in identifying spatially variable genes (SVGs) and segmenting spatial domains across various spatial resolutions, from multicellular to cellular levels. The PROST Index can prioritize spatial expression variations, facilitating biological insights. The framework is validated on datasets from different platforms and resolutions, showing robustness and adaptability. PROST's ability to integrate spatial information and gene expression profiles makes it a flexible and powerful tool for analyzing spatial transcriptomic data.