Osteopenia, also known as low bone mass or low bone density, is a disorder characterised by low bone mineral density, which can be viewed as the precursor to osteoporosis. In contrast to the subjectivity of doctors and the necessity for considerable professional training, computer vision does not present similar issues in its applications in the medical domain and has thus been a topic of great interest. In this paper, we propose a Siamese-based Network for detecting paediatric wrist osteopenia. More precisely, the features given by the two branches of our model are concatenated and fed as input to fully connected and Softmax layers in order to make the prediction. In addition, several Convolutional Block Attention Modules (CBAM) were integrated to improve the feature extraction performance. Experimental results using the GRAZPEDWRI-DX dataset indicate that the proposed method has promising potential to detect wrist osteopenia.