A quantum image saliency detection scheme is proposed, which is more efficient than the traditional methods. In order to represent and store RGB images in quantum computers and calculate the contrast between different pixels, the scheme adopts three-qubits to describe color information. The 22n×2 image matrix is encoded as a quantum superposition state. Combined with Hadamard gate and controlled rotation operator, the global color contrast of pixels in RGB three-channels can be reflected by computing probability amplitude of basis states. The normalized color contrast and location information is achieved using a finite number of projective measurements, and the saliency map is constructed. The realization of related quantum circuits and complexity analysis are presented. The contrast experiments are carried out with several traditional saliency detection algorithms. Results show that the proposed scheme has good detection effect and higher detection efficiency.