[英]How can I draw the Confusion Matrix when using image_dataset_from_directory in Tensorflow2.x?
My TF version is 2.9 and Python 3.8.我的 TF 版本是 2.9 和 Python 3.8。
I have built an image binary classification CNN model and I am trying to get a confusion matrix.我已经建立了一个图像二进制分类 CNN model 并且我正在尝试获得一个混淆矩阵。
The dataset structure is as follows.数据集结构如下。
train/
│------ benign/
│------ normal/
test/
│------ benign/
│------ normal/
The dataset configuration is as follows.数据集配置如下。
train_ds = tf.keras.utils.image_dataset_from_directory(
directory = train_data_dir,
labels="inferred",
validation_split=0.2,
subset="training",
seed=1337,
color_mode='grayscale',
image_size=(img_height, img_width),
batch_size=batch_size,
)
val_ds = tf.keras.utils.image_dataset_from_directory(
directory = train_data_dir,
labels="inferred",
validation_split=0.2,
subset="validation",
seed=1337,
color_mode='grayscale',
image_size=(img_height, img_width),
batch_size=batch_size,
)
test_ds = tf.keras.utils.image_dataset_from_directory(
directory = test_data_dir,
color_mode='grayscale',
seed=1337,
image_size=(img_height, img_width),
batch_size=batch_size,
)
I wrote the code referring to the following link to get the confusion matrix.我参考以下链接编写了代码以获取混淆矩阵。 Reference Page 参考页
And this is my code about the confusion matrix.这是我关于混淆矩阵的代码。
predictions = model.predict(test_ds)
y_pred = []
y_true = []
# iterate over the dataset
for image_batch, label_batch in test_ds: # use dataset.unbatch() with repeat
# append true labels
y_true.append(label_batch)
# compute predictions
preds = model.predict(image_batch)
# append predicted labels
y_pred.append(np.argmax(preds, axis = - 1))
# convert the true and predicted labels into tensors
true_labels = tf.concat([item for item in y_true], axis = 0)
predicted_labels = tf.concat([item for item in y_pred], axis = 0)
from sklearn.metrics import confusion_matrix
cm = confusion_matrix(true_labels, predicted_labels)
print(cm)
y_pred and y_true were obtained from test_ds as above, and the results of confusion matrix were as follows. y_pred 和 y_true 是从上面的 test_ds 中得到的,混淆矩阵的结果如下。
[[200 0]
[200 0]]
So I tried outputting true_labels and predicted_labels, and confirmed that predicted_labels are both 0 as follows.所以我尝试输出true_labels和predicted_labels,并确认predicted_labels都是0,如下所示。
print(true_labels)
<tf.Tensor: shape=(400,), dtype=int32, numpy=
array([0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1, 0,
1, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 0, 0, 1, 1, 0,
0, 0, 1, 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 0,
0, 1, 1, 0, 0, 0, 1, 0, 1, 1, 1, 1, 0, 0, 0, 1, 1, 0, 1, 0, 0, 0,
1, 1, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0,
0, 0, 0, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 1, 0, 1,
0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1,
1, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 0, 0, 1,
1, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 1,
0, 1, 1, 1, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0,
1, 1, 0, 1, 0, 0, 1, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0,
0, 0, 1, 0, 1, 0, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 0, 1, 0, 1, 0, 1,
0, 0, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0,
1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 0, 0,
1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 1, 0,
0, 0, 1, 0, 1, 0, 0, 0, 1, 1, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0, 1, 1,
1, 0, 1, 0, 1, 0, 1, 0, 1, 1, 0, 1, 1, 0, 1, 0, 1, 1, 0, 0, 0, 0,
0, 1, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0,
0, 0, 1, 1])>
print(predicted_labels)
<tf.Tensor: shape=(400,), dtype=int64, numpy=
array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0], dtype=int64)>
I'm not sure why predicted_labels are all zero.我不确定为什么 predict_labels 都为零。
But this is wrong.但这是错误的。 I think the following results are correct.我认为以下结果是正确的。
[[200 0]
[0 200]]
What is wrong?怎么了? I've been struggling for a few days.我已经挣扎了几天。 Please please help me.请帮帮我。
Thanks a lot.非常感谢。
In case of Image Binary Classification, threshold should be used to obtain predict label after model.predict(test_ds).在图像二元分类的情况下,应使用阈值在 model.predict(test_ds) 之后获得预测 label。 I found that modifying the code in my question y_pred.append(np.argmax(preds, axis = - 1))
to y_pred.append(np.where(preds > threshold, 1, 0))
solved the problem.我发现将问题y_pred.append(np.argmax(preds, axis = - 1))
中的代码修改为y_pred.append(np.where(preds > threshold, 1, 0))
解决了这个问题。 Hope it was helpful to someone.希望它对某人有帮助。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.