Search papers, labs, and topics across Lattice.
3
0
4
HPR-SAM outperforms existing methods by effectively capturing complex anatomical representations, achieving state-of-the-art results in medical image segmentation without the need for prompts.
Image compression, a seemingly benign process, can drastically amplify the power of adversarial attacks, making your image classifiers far more vulnerable than you thought.
Text-to-image safety filters are surprisingly easy to bypass: simple prompt reframing techniques achieve a 74% success rate in generating restricted imagery.
