About visualizing data

1

Hi every one
I need help to visualize a frames in data set when applying each transformation operation?

    mean=[0.485, 0.456, 0.406]
    std=[0.229, 0.224, 0.225]
    normalize = Normalize(mean=mean, std=std)
    spatial_transform = transforms.Compose([transforms.RandomRotation(20),
                                            transforms.RandomResizedCrop(224),
                                            transforms.RandomHorizontalFlip(),
                                            transforms.ColorJitter(hue=.05, saturation=.05),
                                            transforms.ToTensor(),
                                            transforms.Normalize([0.485, 0.456, 0.406],[0.229, 0.224, 0.225])])


  
    vidSeqTrain = makeDataset(trainDataset, trainLabels, spatial_transform=spatial_transform, seqLen=seqLen)
    
        
    vidSeqTest = makeDataset(data2, label2, seqLen=seqLen, spatial_transform= spatial_transform)
    
 
    testLoader = torch.utils.data.DataLoader(vidSeqTest, batch_size=trainBatchSize,
                            shuffle=True, num_workers=int(numWorkers/2), pin_memory=True)
    
    # torch iterator to give data in batches of specified size
    trainLoader = torch.utils.data.DataLoader(vidSeqTrain, batch_size=trainBatchSize,
                            shuffle=True, num_workers=numWorkers, pin_memory=True, drop_last=True)
2

If you want to visualize the PIL.Image after each transformation, you could create a custom transformation and call e.g. img.show() (or use any other library to plot the image):

class Visualize(object):
    def __call__(self, img):
        img.show()
        return img
        
spatial_transform = transforms.Compose([transforms.RandomRotation(20),
                                        Visualize(),
                                        transforms.RandomResizedCrop(224),
                                        Visualize(),
                                        transforms.RandomHorizontalFlip(),
                                        Visualize(),
                                        transforms.ColorJitter(hue=.05, saturation=.05),
                                        Visualize(),
                                        transforms.ToTensor(),
                                        transforms.Normalize([0.485, 0.456, 0.406],[0.229, 0.224, 0.225])])


to_pil_image = transforms.ToPILImage()
x = torch.randn(3, 256, 256)
img = to_pil_image(x)
out = spatial_transform(img)
1 Like