<p>Sequential monitoring of image data is an emerging research area in statistics and machine learning, which has various applications in satellite imaging, medical diagnostics, manufacturing industries, etc. The problem of image monitoring is challenging mainly because the images are often not geometrically aligned, and hence, the problem of rigid-body image registration is also inherently associated with it. Most methods in the literature require image registration as a pre-processing step. To avoid this scenario, we propose a shape and size based image object monitoring algorithm which is invariant to rotation and translation of the image object, and thus capable of disregarding the changes due to rigid-body image transformation. For comparing the boundaries of an image object in two images, we construct a test statistic based on the distribution of radial distances, and propose a CUSUM type control statistic based on that. The primary advantage of the proposed method is that the complications associated with the performance of image registration do not arise. This paper makes an effort to bridge the gap between the domains of statistical shape analysis and image monitoring. Theoretical justifications and numerical studies show that the proposed method works well in many applications.</p>