CVE-2022-41888 TensorFlow is an open source platform for machine learning. When using GPU, `tf.image.generate_bounding_box_proposals` receives a `scores` input that is not checked >

CVE-2022-41888 TensorFlow is an open source platform for machine learning. When using GPU, `tf.image.generate_bounding_box_proposals` receives a `scores` input that is not checked

>

In TensorFlow 2.10.1, 2.9.3, and TensorFlow 2.8.4, the `tf.image.generate_bounding_box_proposals` function does not check for rank 4 scores. This issue can lead to an OOM when a large number of images (for example, photos) are processed. To fix this, we have patched the `tf.image.generate_bounding_box_proposals` function. This issue has been reported on TensorFlow GitHub issue tracker. We will assign this issue a medium severity and patch it in TensorFlow 2.10.1, 2.9.3, and 2.8.4. This issue has been reported on TensorFlow GitHub issues tracker. We will assign this issue a high severity and patch it in TensorFlow 2.10.1, 2.9.3, and 2.8.4. We will also cherrypick this commit on TensorFlow 2.10.1, 2.9.3, and TensorFlow 2.8.4, as these are also affected and still in supported range. In TensorFlow 2.10.1, 2.9.3, and TensorFlow 2.8.4, the `tf.image.generate_bounding_box_proposals` function does not check for rank 4 scores. This issue can lead to an

TensorFlow 2.10.1

, 2.9.3, and TensorFlow 2.8.4
In TensorFlow 2.10.1, 2.9.3, and TensorFlow 2.8.4, the `tf.image.generate_bounding_box_proposals` function does not check for rank 4 scores was originally submitted on GitHub by finkanmanus in the issue tracker at https://github.com/tensorflow/tensorflow/issues/6881 as:
TensorFlow 2: Generate Bounding Box Proposal Function Does Not Check for Rank 4 Scores

The following code is truncated to fit on this page without sacrificing readability:

# function to generate bounding box proposals based on l2 loss with tf-idf # -------------- # use -1 to force a bounding box proposal from the entire image # --------------- def generate_bounding_box_proposals(image, score = -1): """Generates a list of bounding boxes that enclose the given image and returns an array of `tf-idf` scores.""" numImages = tf._global_variables()["num_images"] numProposals = tf._global_variables()["num_proposals"] nbPairs = tf._global_variables()["nbPairs"] ids = np . arange ( numImages ) . reshape (( num

References

Subscribe to CVE.news
Don’t miss out on the latest issues. Sign up now to get access to the library of members-only issues.
jamie@example.com
Subscribe