SiamVIT. The upper and lower branches are both VIT structures. The upper branch processes templates containing location information. Its number of feature embeddings is equal to the number of templates. The lower branch processes γ-ray map information, and its number of feature embeddings equals the number of patches in the γ-ray map. The inputs of the matching block are the feature embeddings of the template and the γ-ray map, and the output is a matrix. A patch corresponding to high values in the matrix may contain a point source, and the position of the point source is the same as the position of the point source in the template corresponding to the high values.