Multi-modality medical images

Multi-Modality Medical Images Feature Analysis... 

Keywords— Multi-modality medical images, feature analysis, shape, color and texture descriptor. 

Fig. 4 Example of multi-modality medical image taken from ImageCLEF...

Figure 1,2 and 3 shows the accuracy value of precision and recal for texture, shape and color descriptor. It shows that CT, XR, US, PET and PX are suitable to be classified using texture descriptor as depicted in Fig. 1. As for moment shape descriptor, the result almost similar with texture but the value of accuracy is higher as shown in Fig. 2. It explains that shape descriptor using local level has better presentation compare to texture descriptor in analysing multi-modality medical images. Finally the color descriptor in Fig.3 shows that GX, PX, NM and PET have higher value. This is due to both modalities have used more colors compare to other modalities which concentrate only on grey-scale image.

---