📝 PDiscoNet: Semantically Consistent Part Discovery for Fine-Grained Recognition 🔭

"Proposes an end-to-end trainable method to discover object parts by using only image-level class labels along with priors encouraging the parts to be discriminative, compact, distinct from each other, equivariant to rigid transforms, and active in at least some of the images." [gal30b+] 🤖 #CV

⚙️ https://github.com/robertdvdk/part_detection
🔗 https://arxiv.org/abs/2309.03173v1 #arxiv

https://creative.ai/system/media_attachments/files/111/028/377/448/872/011/original/d5f3537f09610fb6.jpg

https://creative.ai/system/media_attachments/files/111/028/377/509/948/933/original/e4e32bba1a37c5de.jpg

https://creative.ai/system/media_attachments/files/111/028/377/577/888/338/original/98a4081a71fc9cc3.jpg

https://creative.ai/system/media_attachments/files/111/028/377/717/058/933/original/658ea67f0d5b7dbf.jpg