图像排序张量流

Question

我正在使用张量流运行滑动窗口的代码。

这是为了Theano，这就是我得到图像尺寸排序错误的原因。

任何人都可以告诉我如何修改张量流量？

错误：

ValueError：检查时出错：预期input_2具有形状（无，224,224,3）但是具有形状的数组（1,3,224,224）

image=load_img('\e.jpg')
image= np.array(image).reshape((3264,5896,3))
image = image.astype('float32')
image /= 255
plt.imshow(image)
print(image.shape)


#%%SLIDING WINDOW
 ''
def find_a_object(image, step=224, window_sizes=[224]):
    boxCrack = 0  
    locations = []
    for win_size in window_sizes:
        #top = Y, left = X
        for Y in range(0, image.shape[0] - win_size + 1, step):
            for X in range(0, image.shape[1] - win_size + 1, step):
                # compute the (top, left, bottom, right) of the bounding box
                box = (Y, X, Y + win_size, X + win_size)

                #crop original image 
                cropped_img = image[box[0]:box[2], box[1]:box[3]]
                #reshape cropped image by window
                cropped_img = np.array(cropped_img).reshape((1,3,224,224))


                boxCrack = predict_function(cropped_img)

                if boxCrack ==0:
                    #print('box classified as crack') 
                    #save location of it                 
                    locations.append(box)
                    print("found")
        return locations
#%%FUNCTIONS
def predict_function(x):
    result =  model.predict_classes(x)
    if result==0:
        return 0
    else:
        return 1   
##SHOW CROPPED IMAGE
#def show_image(im):
#    plt.imshow(im.reshape((100,100,3)))
)

#DRAW BOXES FROM LOCATIONS when classified
def draw_boxes(image, locations):
    fix,ax = plt.subplots(1)
    ax.imshow(image)       
    for l in locations:
        print (l)
        rectR = patches.Rectangle((l[1],l[0]),224,224,linewidth=1,edgecolor='R',facecolor='none'
        ax.add_patch(rectR)


#%%get locations from image
locations = find_a_object(image)

draw_boxes(image,locations)

Answer 1

好吧，错误信息非常清楚，TF中的排序应该是(None, dim_x, dim,_y, n_channels)。我认为如下所示转置图像数组应该可以解决问题。

cropped_img = image[box[0]:box[2], box[1]:box[3]]
#reshape cropped image by window
cropped_img = np.array(cropped_img).reshape((1,3,224,224))
cropped_img = cropped_img.transpose(0,2,3,1)