数据集图像中的聚类和作物癌症片段

问：

数据集分为 train、test、val，每个都有 7 种癌症亚型。我想做的是对每个图像进行聚类，并保存包含每个图像中癌症图像的重要片段。

该任务看起来像扩展数据集，我正在使用 K-Means-Clustering

data = []
label = []
path = "C:/Users/ergun/Desktop/ergunDosyalar/DERS/bilgisayarprojesi/dataset/test/"

n = 0
pb = 0
udh = 0
fea = 0
adh = 0
dcis = 0
ic = 0
IMG_SIZE = 32
for file in os.listdir(path):
    if file[0]=='0':
        for imgfil in os.listdir(path+file):
            img=cv2.imread(path+file+'/'+imgfil)
            img=cv2.resize(img,(IMG_SIZE,IMG_SIZE))
            img=img.astype('float32')
            n+=1
            label.append("N")
            data.append(img)
    elif file[0]=='1':
        for imgFil in os.listdir(path+file):
            img=cv2.imread(path+file+'/'+imgFil)
            img=cv2.resize(img,(IMG_SIZE,IMG_SIZE))
            img=img.astype('float32')
            pb+=1
            label.append("PB")
            data.append(img)
    elif file[0]=='2':
        for imgfil in os.listdir(path+file):
            img=cv2.imread(path+file+'/'+imgfil)
            img=cv2.resize(img,(IMG_SIZE,IMG_SIZE))
            img=img.astype('float32')
            udh+=1
            label.append("UDH")
            data.append(img)
    elif file[0]=='3':
        for imgFil in os.listdir(path+file):
            img=cv2.imread(path+file+'/'+imgFil)
            img=cv2.resize(img,(IMG_SIZE,IMG_SIZE))
            img=img.astype('float32')
            fea+=1
            label.append("FEA")
            data.append(img)
    elif file[0]=='4':
        for imgFil in os.listdir(path+file):
            img=cv2.imread(path+file+'/'+imgFil)
            img=cv2.resize(img,(IMG_SIZE,IMG_SIZE))
            img=img.astype('float32')
            adh+=1
            label.append("ADH")
            data.append(img)
    elif file[0]=='5':
        for imgfil in os.listdir(path+file):
            img=cv2.imread(path+file+'/'+imgfil)
            img=cv2.resize(img,(IMG_SIZE,IMG_SIZE))
            img=img.astype('float32')
            dcis+=1
            label.append("DCIS")
            data.append(img)
    elif file[0]=='6':
        for imgFil in os.listdir(path+file):
            img=cv2.imread(path+file+'/'+imgFil)
            img=cv2.resize(img,(IMG_SIZE,IMG_SIZE))
            img=img.astype('float32')
            ic+=1
            label.append("IC")
            data.append(img)

data = np.array(data)

data_label = []
for i in label:
    if i=="N": data_label.append(0)
    elif i=="PB" : data_label.append(1)
    elif i=="UDH" : data_label.append(2)
    elif i=="FEA" : data_label.append(3)
    elif i=="ADH" : data_label.append(4)
    elif i=="DCIS" : data_label.append(5)
    elif i=="IC" : data_label.append(6)
data_label = np.array(data_label)

data = data/255.0
reshaped_data = data.reshape(len(data),-1)

kmeans = KMeans(n_clusters=2, random_state=0, n_init='auto')
clusters = kmeans.fit_predict(reshaped_data)

image_cluster = kmeans.cluster_centers_[kmeans.labels_]
#image_cluster = image_cluster.reshape(reshaped_data.shape[0], reshaped_data.shape[1])
plt.imshow(image_cluster)
plt.show()

使用包含癌症的图像段展开数据集