Akdeniz, Ahmet SezerOzgur, BerkanSahin, EmreKaradag, Ozge OztimurGunizi, Ozlem Ceren2026-01-242026-01-242024979-8-3503-8897-8979-8-3503-8896-12165-0608https://doi.org/10.1109/SIU61531.2024.10600996https://hdl.handle.net/20.500.12868/507032nd IEEE Signal Processing and Communications Applications Conference (SIU) -- MAY 15-18, 2024 -- Tarsus Univ Campus, Mersin, TURKEYThis research comprehensively compares two different methodologies for predicting the Ki-67 index- morphological-based analysis and Vision Transformers (ViT). The morphological method focuses on the shape and structural features of tissues and cell structures. On the other hand, Vision Transformers represents an innovative approach developed through the use of attention mechanisms and transformer architectures. ViT offers a different perspective by modeling global context information in recognizing image patterns. This analysis provides a deeper understanding of the accuracy, efficiency, and applicability of current techniques in histopathological image processing, highlighting the potential to advance existing methodologies used in cancer diagnosis. This comparative study aims to evaluate the performance differences between morphological analyses and transformer-based models, identifying the most effective and reliable methods for predicting the Ki-67 index. Experimental analysis, revealed that due to the limited number of labeled data on this domain, traditional morphologic approaches are currently more promising than the vision transformers.trinfo:eu-repo/semantics/closedAccessKi-67ViTMorphologySegmentationCancerConference Object10.1109/SIU61531.2024.106009962-s2.0-85200881597N/AWOS:001297894700218N/A