Bone cancer kills most people worldwide. X-rays, MRIs, and CT scans aid bone cancer diagnosis. Manual processes take too much time and knowledge. Thus, a machine learning system must distinguish malignant bone from healthy bone. Cancer-damaged bones feel different. Cancer and healthy bone scans were similar. Classifying them is hard. Despite continuous innovation in the area of cancer research, it still remains one of the most fatal diseases in the world. It is paramount to harness all resources to bring breakthroughs in cancer treatment. Detection at the initial stage would be a significant stride in prolonging the life of the patient and decreasing the mortality rate. Hence, it is imperative to develop techniques that are innovative, efficient, and with lesser undesirable effects. This paper involves the study of all elements of bone cancer and features to evaluate/estimate its type. ACM is one of the most well-known techniques for segmenting a bone image into individual parts and the random forest method employs a clustering strategy with a maximum of five nodes. Insight into the organ's volumes and contours from an image is seen as highly challenging. It's a collection of operations including filtering out background noise, dividing up the data, mining for features, and making a final pick. It detaches the pixels from the edges of the object. They are treated further to bring about the intended effect. The achieved results are satisfactory, and this method is among the best ones for segmenting greyscale images currently available.