Object alignment for a moving object in the two consecutiely acquired images

I have two images acquired from a moving object. The object is moving in a fixed direction (from left to right in the image) in the scene under controlled illumination conditions. The light is spatially patterned. Here below are the two images acquired:

The two images look almost identical to each other, but they actually not, because the object is moving during image acquisition (background is totally the same).

What I want to do is to align the position of the object in the two images, that is to say, estimate the actual spatial shift of the object in pixel unit for the second image (right) compared to the object in the first image (left).

My original though is to segment the object, get its mask in the image, and to estimate the spatial offset of corresponding centroids for the object masks. But this method does not work quite well, since accurate segmentation of the object is not easy for the patterned images. Is there any other possible solutions to the problems?