Shoulder injuries, especially those caused by rotator cuff tears, are a pervasive problem in human and canine patients. The ability to quickly and accurately diagnose rotator cuff tears is important for early and targeted treatment. Prior clinical work has demonstrated that ultrasound and magnetic resonance imaging effectively diagnose these tears. However, there has been limited use of quantitative image processing to validate the imaging and clinical findings. The aim of our study was to evaluate the use of medical image processing to objectively assess tendon injuries in ultrasound and magnetic resonance images of biceps and supraspinatus tendons of dogs. Histogram peak intensity measurements were performed on regions of interest drawn on ultrasound and magnetic resonance images of the supraspinatus and biceps muscles and tendons in 5 dogs with known front limb lameness and pain localized to the shoulder joints. Peak intensity increased from untorn supraspinatus and biceps tendon to torn tendon at a greater magnitude in ultrasound imaging than in magnetic resonance imaging. The histogram analysis suggested that ultrasound imaging is a superior method for diagnosing torn supraspinatus and biceps tendons based on contrast. Furthermore, this conclusion is strengthened by practical considerations that favor ultrasound imaging as a primary form of diagnosis.