a device body;

a supporting portion being provided within an inner side of said device body and accommodating a moisture absorption member in such a manner to allow an excrement of the living creature to be left on said moisture absorption member, the moisture absorption member is elected from a group consisting of non-woven fabric, sand grains, corks or a combination among the non-woven fabric, the sand grain and the corks;

an image module, arranged at said device body for dynamically capturing one or more images of the excrement in said supporting portion and outputting an image information of the excrement; and

a first analysis module arranged in said device body and connected with said image module, to analyze and calculate a defecation mode with the image information based on preset or accumulated data, so as to generate a signal when an abnormal defecation mode is diagnosed;

wherein a coordination of the supporting portion, the image module, and the first analysis module conduct a multi-laver operation of artificial intelligence based on a framework of convolution neural network, wherein when the excrement captured in an image, from the one or more images, is covered by sand gravel, sequentially capture image blocks of various positions and utilize the multi-layer operation of artificial intelligence to restore images, for the one or more images, of uncovered excrement so as for a comparison to images of internal or external databases for calculation and analysis of the abnormal defecation mode.