Quantum TV algorithm, the result of which only takes values from a finite discrete set(the quanta set), is a kind of quantized TV regularized algorithm. It can find applications mainly in binary image restoration, image quantization and image segmentation. Different part of an image varies differently, thus it needs varied number of quantized levels. Different image histogram also corresponds to different quantized levels. We made improvements to the quantum TV algorithm according to the above statements. Experiments and related data demonstrated that the improved algorithm showed better appliance in image quantization.