sassoftware / sas-viya-machine-learning Goto Github PK

Hello,

I have a question that hopefully a member from this project can help answer.

I'm using the image action set to load and reshape some images. I borrowed the imageShow() function from @pepecura's Jupyter notebook Car-Damage-Image-Analysis-for-Insurance.ipynb

# define the function to display the processed image files.
def imageShow(session, casTable, imageId, nimages):
    a = session.table.fetch(sastypes=False,sortby=[{'name':'_id_'}],table=casTable,to=nimages)
    fig = plt.figure(figsize=(20, 20))
    for i in range(nimages):
        imageData = a['Fetch'].iloc[i][imageId]
        img_np = cv2.imdecode(np.fromstring( imageData, np.uint8),1)
        fig.add_subplot(1,nimages,i+1)
        plt.imshow(img_np)
        img_np[:,:,[0,2]]=img_np[:,:,[2,0]]
        plt.xticks([]), plt.yticks([])

The issue I'm having is that when I use it to display my images, they all show up overly blue:

Source images are .jpg and if open any of them with a regular image viewer they show up with normal colors:

Could you please help me understand why my pictures are showing up with blue tinting or recommend a better function to display images loaded into a CAS table?

Thanks!